Continuing Professional Development Report

Overall summary

This document contains a record of my Continuous Professional Development activities.

CPD should be a key part of a professional statistician’s work ethic. CPD means the development of the “whole person” statistical and non-statistical, technical and personal. CPD evaluation is essentially a self-assessment process.

Each year, I should have at least 60 hours of learning hours, across at least three of the following pre-defined categories.

• Work based learning
• Professional activity
• Formal / educational
• Self-directed learning
• Other

Year	Learning Hours
2023	67.0	✅
2022	60.0	✅
2021	63.0	✅
2020	69.5	✅
2019	100.0	✅
2018	61.5	✅
2017	62.5	✅

Detailed Report: 2023

In 2023 I completed 67 hours of CPD training, across 4 categories.

	Activity Type	Learning Hours
	Professional activity	31
	Work based learning	20
	Formal / educational	9
	Self-directed learning	7
Total	—	67.00

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
Data meetups	5	I attended a couple of NEDS, Databrick and Lead Dev meetups in 2023. Databricks is totally new to me, and I found it really useful to attend some introductory workshops. The Lead Dev panel on Burnout in tech was really useful in helping me support my team and also led to me recognising that I was burnt out myself.	Recognising the symptoms in the burnout session helped me take the time I needed to start recovering and get help.	Databricks is a tool that clients often ask for help with. I now feel more confident getting started and helping them decide if it's useful for their needs.
Emerging Leaders Year 2	10	I'm on a two year leadership programme with Vistage. The second year included topics on service management, managing change and emotional intelligence.	I found the emotional intelligence session hugely enlightning. I am not more aware of my strengths, and I check in to make sure I am utilising them wisely.	I'm also more aware of when my strengths can get in my way, and look out for signs that I might be communicating poorly, letting my overachieving get in the way or being too commanding.
Jumping Rivers internal tech talks	12	Weekly internal knowledge sharing sessions. Some technical topics like powerpoints with {officer}, Quarto tempaltes and git interactive rebasing There were also sessions on specific Jumping Rivers knowledge, such as how we manage our services	This kept me informed with latest working practices within Jumping Rivers, and tools that others had found useful within their client work.	Having clear, well defined working practices that we all follow allows Jumping Rivers to deliver quality service.
Large analysis of patient record data	6	I was involved in a very large project analysing patient records for a specific disease. I had to learn more about the specific disease, and the appropriate analysis techniques. I also had to design a workflow that would allow our German analyst to run the analysis seperately and then allow use to recombine the data at a later date.	This was a complex project and I learned a lot about how to handle a large codebase which deals with sensitive data.	The experience gained in this project has directly influenced the way I approach new programmatic statistical analysis projects.
RWeekly	5	RWeekly is a newsletter showcasing the lastest blog posts and package updates in the RCommunity. I regularly read the highlighted posts and investigated new relevant R packages.	The programming world moves quickly, and following this newsletter helps me stay up-to-date.	As a professional programmer, clients expect me to know the latest R packages and best practices.
Shiny bookclub	2	Russ organised an internal bookclub in work for Shiny. We talked about SASS, shinytest2 and did a review of a shiny application.	Building efficient, robust, professional shiny applications is a key part of my work at Jumping Rivers. Russ taught me some tools which will be very useful in future projects.	We also discussed and critqued user interfaces for shiny applications which is a key aspect for our end users.
Significance Magazine	2	I get Signficance magazine as part of my RSS membership. I find that the short, pop-science style articles are a great way for me to keep my statistical knowledge fresh.	As I do more programming at the moment, I sometimes worry that my general statistical knowledge will get rusty. These articles are perfectly pitched to remind me of key modelling approaches.	Having a broad range of statistical tools means I'm more likely to suggest a suitable model for my clients.
Recontres R	2	R conference for French speakers. I was an invited keynote.	Some interesting talks here, in particular the one from Aurélie on importer syndrome. Attending a conference in a different language gave me a new perspective on how to make content inclusive for people who are less comfortable in the language of the conference.	There were a number of talks on reproducibility. A number of my clients work in government and are keen to create more "RAP" workflows.
Autism Open University course	6	I work with many people who are neurodiverse and so I wanted to learn a little more about Autism. I want to make sure that I an inclusive, and that as a leader and educator, I create a welcoming environment. This was an 8 week course, which I studied in my own time over a couple of months.	I didn't know much about Autism before starting this course. By the end, I had learned more about history of Autism, the variety of the spectrum and had a greater awareness of some way I could support neurodiverse people.	This course has helped me understand the different needs of autistic people, which has lead to me being more considerate when designing courses and working practices.
RSS Conference Harrogate	10	I presented a workshop on spatial data, a talk on refactoring code and a workshop on data science portfolios.	The RSS conference provides a great way for me to keep up to date with developments in the statistical world, particulary in academia.	Being relevant within the statistical world means I have a large tool box of techniques to use, to find the right approach for my clients.
NICD Data Innovation Showcase	2	I gave a talk introducting Quarto to a non technical audience. I created an example based on healthcare.	There were some great introductory talks on statistical methods, version control and data science pipelines. I reflected on how I could better introduce those topics when I'm talking with non-technical audiences.	Being able to explain the benefits of data science workflows and techniques to non-technical people is a key part of my work as a consultant.
Responsive Web Applications	3	A formal workshop taught by Dr Keith Newman. I learned what makes web applications frustrating to use, and tools to make the applications I build more responsive to the user.	I regularly build web applications but I didn't really understand about HTML and JS. Keith helped me understand how the design choices I make will affect my users.	I will use the skills I learnt to make more responsive web applications for my users.
Shiny in Production conference	2	A very niche, but relevant conference for me. I learned so much from Cara's keynote on plot styling. Anna's talk on "Why users don't use your app" was another useful one.	As a Shiny expert, it's important for me to stay up-to-date with the latest developments in shiny.	This gave me a few ideas for technical approaches to shiny applciations, but more importantly UI/UX elements which I will bring into future projects.

Detailed Report: 2022

In 2022 I completed 60 hours of CPD training, across 4 categories.

	Activity Type	Learning Hours
	Professional activity	25
	Formal / educational	12
	Work based learning	12
	Self-directed learning	11
Total	—	60.00

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
Emerging Leaders Year 1	12	I'm on a two year leadership programme with Vistage. The first year included topics on finance, performance management and delegation.	Thinking about delegation, lead me to reconsider what aspects are within my control and what are not. I realised that I can be guilty of both doing work for those junior to me, but also trying to do the thinking of those more senior to me (and feeling frustated). I've started applying more focus to my efforts, utilising my skills and thought leadership at the level in which I can have the most impact.	As a manager of a small team, the skills I learned in the performance management and delegation have helped me ensure that my reports have automony and purpose. I also have taken the financial planning ideas into my project budget.
Local data meetups	4	I attended a number of R-Ladies and PyData Lancaster meetups. I also presented at an R-Ladies/Py-Ladies cross-over event. Topics varied, but included "my first data science project" and "Introduction to ML Ops"	These meetups often have broad topics, and it opens me up to new areas I might not have considered. I also believe it's important to support local meetups, and it's a great way to make connections.	New tools and tips ensure that the work I deliver to clients is using the latest best practice.
RWeekly	3	RWeekly is a newsletter showcasing the lastest blog posts and package updates in the RCommunity. I regularly read the highlighted posts and investigated new relevant R packages.	The programming world moves quickly, and following this newsletter helps me stay up-to-date.	As a professional programmer, clients expect me to know the latest R packages and best practices.
Significance Magazine	2	I get Signficance magazine as part of my RSS membership. I find that the short, pop-science style articles are a great way for me to keep my statistical knowledge fresh.	As I do more programming at the moment, I sometimes worry that my general statistical knowledge will get rusty. These articles are perfectly pitched to remind me of key modelling approaches.	Having a broad range of statistical tools means I'm more likely to suggest a suitable model for my clients.
Appsilon shiny conference (virtual)	3	As a shiny expert, it's important for me to keep up to date with the latest developments. Highlights for me included the Panel discussion "Women in Shiny" and "Testing Shiny with {testthat}"	There were a number of technical talks which gave me new tools to help make my shiny applications faster and more professional.	I design and build shiny applications for a number of clients. The skills learnt will allow me to deliver more professional tools.
Posit Conference (virtual)	6	There were some major updates announced at this conferences. I watched a couple of the talks, and went away to learn more about the tools afterwards. The two big announcements were Quarto & Shiny for Python.	Quarto as RMarkdown v2 solves a number of problems I've had an an educator, particularly across multiple languages.	Quarto is a modern way for automating reports. With this tool, I'll be able to deliver reproducbile reports to my clients.
Quarto	6	Quarto is the new tool for automated reporting. As this is a key part of my work, I took time in August to learn more about the features in depth. This then led into workshop that I delivered at the RSS conference in September.	Quarto is a tool I now use weekly to write reports and presentations. Taking the time to learn the new features in depth helped me quickly transition from R Markdown to Quaro.	Quarto is a tool that my clients will be asking about soon for writing statistical reports.
RSS Conference Aberdeen	6	I presented a workshop on Quarto, a talk on {plumber} and contributed to a workshop on communicating statistics. I also attended a number of talks.	The RSS conference provides a great way for me to keep up to date with developments in the statistical world, particulary in academia.	Being relevant within the statistical world means I have a large tool box of techniques to use, to find the right approach for my clients.
Shiny in Production	6	This conference was organised by my work and brought together many technical experts. Highlights included, Colin Fay's destroy all widgets which focused on the userbility of applications, and Caterina's summary of web app tools. Mike's talk {pins} was also very informative.	As a Shiny expert, it's important for me to stay up-to-date with the latest developments in shiny.	This gave me a few ideas for technical approaches to shiny applciations, but more importantly UI/UX elements which I will bring into future projects.
Web Accessibility	6	I developed an app for a client and they wanted me to ensure the app met WGAC 2.1 standard. I learned about the standards, and how to implement things like skip links in Shiny applications. I also learned the importance of correct header ordering.	I was aware of some basics around web accessibility, but this project forced me to delve deeper into the details of the standard. I also learned some tools which can help me check my apps in future e.g. Lighthouse	As a developer of user tools, it's important I build tools which are accessible for all.
Interrupted Time Series	3	ITS analysis was required for a project at work. I had not done this before. Effectively it involves looking at statistical properties of time series before and after an "interruption". I researched about the approach, in particular the assumptions about sample size.	ITS is another analysis in my toolbox that I can use.	ITS is often used in official statistics analysis. It's also quite simple to understand and interpret.
How to write a statistical analysis plan	3	I was the lead author on a Statistical Analysis Plan for a large project at work. I had never written a SAP before and didn't know what one should contain. I contacted a	A SAP is a great way to plan the statistical analysis before the work begins. There is a specific structure required which I have now become more familiar with.	This is an essential skill for working on projects with pharma.

Detailed Report: 2021

In 2021 I completed 63 hours of CPD training, across 3 categories.

	Activity Type	Learning Hours
	Professional activity	28
	Work based learning	26
	Self-directed learning	9
Total	—	63.00

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
Writing a time series course	8	I wrote the material for a one-day time series course. In order to do this, I did a large amount of research both into statistical models for time series, and their implementation in R. I read the book Forecasting: Principles and Practice, which helped me learn more about the R packages {fable}, {feasts} and {tsibble}. I also refreshed my knowledge on Exponential Smoothing and ARIMA models.	Having a refreshed knowledge of important time series models such as ETS and ARIMA will help me work on projects which require forecasting. I've become more confident using the {tidyverts} ecosystem, in particular visualising time series data with {feasts}.	Many of our clients work with time series data. Having a good understanding of both the statistical methods and the implementation in R will allow me to meet their needs.
Introduction to Publishing	1	A webinar on what it's like to write/publish/translate a statistical/programming book in the early 2020s.	I've previously been asked to contribute to writing a textbook. This session gave me a good understanding of what the process is like and helped me think about the best way to deliver material in a formal text book.	Writing a book is a great way to support the community, it might also allow me to deliver training material in a more formal and accessible way.
RSS Data Science with Andrew Ng	1	The RSS Data Science group hosted a fireside chat with Andrew Ng. Andrew discussed the importance of AI leads having both deep technical & business knowledge. He also highlighted that ML engineers would benefit from more exploratory data analysis skills and how sometimes it's better to just fix your model, and focus on using the data in the best way.	Andrew got me thinking about the importance of maximising the usefulness of the data, rather than getting to fixed on model tweaks. He also reminded me that sometimes you just need to try something and if it doesn't work at all, thats just part of the process. I'll keep these principles in mind during my next modelling consultancy project.	By focusing on utilising the data on a couple of well selected models rather than just throwing different models at the data, I should be able to build more appropriate, interpretable models for my client.
Learning about {golem}	6	I am exploring a number of different frameworks in which to create shiny applications in a robust way. A popular framework is using the {golem} package to deploy shiny applications from within an R package. I watched a number of talks on {golem} and read the book Engineering Production-Grade Shiny Apps	I'm a confident shiny developer but still don't feel like I work in a way that meets best practices. Reading about this approach has helped clarify the ways in which I can develop robust shiny applications in a more professional manner. I will also start using the "documentation first" approach detailed in the book. This keeps UI and server components completely separate until necessary.	As part of my consultancy work, I regularly develop shiny applications for clients. By using the {golem} infrastructure, I'll be able to develop more robust, well-structure applications for my clients.
Shiny Developer Series	3	I've been watching the shiny developer series (18 episodes). These are interviews with shiny developers showcasing the tools they use and approaches they take during development. Favourite episodes include Dean Attali, Colin Fay and Nick Strayer.	I would like to become a more advanced shiny developer. In order to do this I need to learn the different tools available for shiny, and how they can integrate with other technologies. The interviews are a great way for me to get an overview of the additional packages, whilst the demonstrations give me a more hands support.	Most of the client projects that I work on involve developing shiny applications. As an advanced shiny developer, I'll be able to develop more robust and efficient applications for my clients.
EdinbR Meetup Shiny Nirvana	1	I attended the virtual EdinbR meetup to hear Eric Nantz talk about Shiny.	Eric's talk highlighted some key resources which I can use in my journey to become an advanced shiny developer. In particular, he discussed the best practices from the Mastering Shiny book, Shiny App Stories and Outstanding User Interfaces with Shiny. I haven't done much work with development for mobile shiny applications, so {shinymobile} should also be useful.	As a shiny developer it's important that I follow best practices in order to create robust applications for my clients. This talk also made me think about the importance of the design first approach, which I'm keen to use on my next client project.
Hiring effectively	1	Webinar by Steph Locke on why you should hire junior software engineer.	I found it really useful to hear a case study of the hiring process at Nightingale HQ. I reflected on the approach JR took for recent hiring positions and I hope use my new knowledge when we next need to recruit.	By hiring effectively, we can grow the Jumping Rivers team in a way that hopefully retains talent and increases diversity. Hiring excellent people means that we can provide great quality consultants and trainers.
CSS course	10	An internal training course on CSS delivered by Mohamed El Fodil Ihaddaden over 3 days.	CSS is a key langauage related to front-end development. I can use it to improve the style of shiny applications and markdown docs.	Many clients require our products to be styled to a specific requirements. Having CSS skills will help me both develop and understand existing code bases.
Writing Best Practices	8	Wrote a "best practices in coding" course. In order to write the course I had to do a decent amount of research including academic papers, especially the work of Greg Wilson. I also looked at blog posts and workshop materials.	Writing clean, reproducible code is a key part of being a developer. Whilst researching and writing, my own coding standards increased dramatically.	Being able to write clean code is a key part of being a developer.
Writing Feature Engineering	4	Supported colleague Andrew in developing a workshop on Feature Topics included user stories and backlog prioritisation.	Being able to communicate feature requirements in a clear way is key for developing with others.	I can use my new skills in user stories and estimation can help me accurately capture and communicate features requirements.
useR! 2021	2	I attended the Teaching, Automation and Reproducibility session at useR! 2021.	In this session I learned about using R for automatic grading and new developments in the field of Psychology.	An important part of my role at Jumping Rivers is developing infrastructure around teaching programming languagues. This session (which I was also speaking at) gave me some interesting ideas of packages which might be useful both internally and with my clients.
The Managers Path	6	I have been reading "The Manager's Path: A Guide for Tech Leaders Navigating Growth and Change" by Camille Fournier. I've been working through the activies at the end of each chapter, and discussing the contents with other senior data scientists.	My role is starting to transition from a tech-lead to a more formal manager position. This is a new area for me and so I'm keen to learn from the experienced of others. In particular this book has shaped how I think about and approach one-to-one meetings.	If I can look after and support my people better, they should be happier and more efficient in their work for our clients.
RSS Media training	6	Training for the RSS statisticial ambassadors. Part one was focused on the "How" of communicating, and part two focused on the "What?"	From part one, I realised that I use my hands a lot and so should "set the shot" to include my hands. Good reminders to say technical words (and my name!) slowly. Remember they are new words for the listener. PAUSE. Use full stops. It makes you seem more credible. From part two, the main take away was to come up with a 15 second snippet, do the research, and strong start/finish. Again with the pauses - make the interviewer do the work.	Being able to communicate clearly is a huge part of my job, from client work, to delivering online courses.
PyData Lancaster Meetups	3	I attended two PyData meetups in 2021. October had talks by Rebecca Killick and Simon Spavound. November was Naomi Lourie and Charlotte Smith.	Rebecca's talk on R packages made me consider how to build packages which are simple to maintain. Naomi spoke about her first data science project. This helped me reflect as a manager on what projects I can get my people working on, and to encourage them to share their experience of it early.	PyData meetups allow me to stay well connected to both the latest approaches in data science, and networking connections across the North West.
WhyR 2021	3	WhyR online one-day conference. I watched half of the sessions and helped deliver a session on RStudio Connect.	Andrew's talk on R in Production gave me a good overview to having reliable, maintainable infrastructure to support statistical models. Anna's talk on How to be a Shiny developer helped me reflect on the skills that need to be developed in our teams at JR.	When we deliver a statistical model to a client, we don't just handover the model, but also a way to interact with that model. This usually involves providing some sort of infrastructure. It's important for this to be reliable and easy to maintain.

Detailed Report: 2020

In 2020 I completed 69.5 hours of CPD training, across 6 categories.

	Activity Type	Learning Hours
	Self-directed learning	28.0
	Professional activity	17.5
	Formal / Educational	12.0
	Other	9.0
	Work based learning	2.0
	Formal / educational	1.0
Total	—	69.50

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
STOR-i Conference 2020	3.0	Annual conference of the STOR-i CDT. Showcasing statistics and operational research in industrial application.	The STOR-i conference is a great way for me to keep up to date with new research in statistics, especially in cluster analysis. Useful talks included Tom Flowerdew's insights on data science pipelines and Christine Currie's talk on stochastic processes.	The conference is a great way to keep informed on recent statistical development. Being up-to-date with statistical methodology developments broadens the areas in which I can support my clients.
Tidy Time Series Workshop	11.0	At RStudio Conference I was TAing on Rob Hyndman's Tidy Time Series workshop. His course covered 12 hours of material, including manipulating, plotting and forecasting time series.	I have strong R skills and decent experience at time series analysis, however I had not combined the two before. By teaching (and therefore learning) the Tidy Time Series material, I was able to learn the state of the art R libraries used to analyse time series data. Specifically I learnt the tsibble, feasts and fable R packages. I also refreshed some statistical time series analysis theory, including calculation of Auto Correlation Functions and rules of thumb for model fitting.	Having an in depth, up-to-date knowledge of time series analysis in R will allow me to write robust analysis for client data with time series data.I was already able to perform time series analysis in R, however, using the framework defined by the tsibble packages will allow any code I write for clients to integrate well with the tidyverse. This is important to many clients as the tidyverse is seen as a standard coding framework in the R community.
RStudio Conf 2020	4.0	The RStudio conference is an annual global conference for all users of R. As an R expert and an RStudio partner, it's a massively valuable conference from both a learning and a networking perspective. I attended a number of talks and was working on our booth throughout the conference.	In particular, I found the following talks very useful, Ryan Timpe's talk about learning R with humorous side projects, Kevin Ushey's talk on renv, and Jenny Bryan's keynote about debugging. I learnt new skills related to deploying models in a reproducible manner and picked up some great tips about debugging. I left the conference with a long list of new R packages to learn, and practices to start implementing in my working day.	One of the main reasons clients come to us is for advice using R and RStudio products. By being informed and up-to-date with developments in the #RStats community, this means my clients will be confident they are getting a robust solution.
Brand Me Lancaster Career Mentoring Event	0.5	Presentation delivered by Nick Holmes,Sales Director at AMG Group Ltd. He spoke about the importance of developing a personal brand and steps to create one.	I'm trying to grow and improve my professional identity. An important part of this is the brand I choose to portray. Nick encouraged me to think about my target audience, and what aspects make me stand out against the crowd. I believe that developing a strong personal brand will help me connect with potential clients.	Creating a strong personal brand forces me to define my skills, and act professionally. This will in turn provide a clearer, more professional service to my clients. Having a strong brand may also help my clients demonstrate to their stakeholders, the quality and caliber in my work.
R-Ladies Lancaster R Markdown	1.0	I co-lead the R-Ladies meet up. I assisted Rebecca with the workshop aspect of her talk, and then I presented my own talk on the RStudio conference.	As a professional female programmer, I believe it's important to give back to the community. By running these sessions, I have to design educational content at a suitable level. I also have to organise the sessions, and advertise them. R-Ladies is also a great place to connect with like-minded professionals and discuss issues and finding in the R community. It's a great networking opportunity for me. I also learn lots each session, this could be a new R package that I haven't heard of, or simply a better way to explain a problem to an attendee.	By delivering workshops to the community and supporting new learners, my clients can see that open-source development is important to me, that I'm a skilled educator and that I'm a supportive listener. It might be important to clients that I support STEM education programmes for women. This also demonstrates to clients that I am an active member of the North West RStats community.
Continuous Integration research	1.0	In preparation for some talks I'm giving, and my developer role in work, I watched a number of talks on continuous integration strategies on YouTube and read [blogposts](https://medium.com/@mattia.battiston/why-i-love-trunk-based-development-641fcf0b94a0)	Continuous integration is essential for efficient, robust code development. Through these videos I learnt about different git branching strategies, and which ones lend themselves well to CI.	Continuous integration and git branching strategies are a key part of developing functioning code. Having a good understanding of the options will allow me to make better informed decisions when working with clients.
PyData Lancaster February	0.5	I spoke at a PyData Lancaster meetup about continuous integration. The other speaker was James Walkerdine, Co-Founder of Relative Insight. He told us the story of how he started a language analytics and data science company, and how it has changed and grown into a sustainable business.	James gave a great presentation explaining how his data science company grew from a small size to having three international offices. I learnt about how to support the growth of a data science team, and tips on ensuring clients feel supported. I also had some useful discussions with attendees about continuous integration techniques.	The discussions I had with other attendees introduced me to alternative CI solutions to Travis. This will be useful when discussing deployment options with clients. My learning from James's talk will help me foster sustainable growth within my own team, hopefully leading to a happier, more productive team and therefore better output.
R-Ladies Lancaster Introduction to Shiny	1.0	I ran a beginner friendly Shiny tutorial for R-Ladies Lancaster. I wrote a practical for participants.	Preparing for this workshop made me revisit what it's like to learn shiny as a beginner. I thought about how to explain the different components, and how they all interact.	I regularly have to explain to clients what a shiny app is and how it works. This session helped me practice explaining these complex topics in understandable terms.
RStudio webinar on teaching online	1.0	RStudio Webinar for certified trainers on Teaching online on short notice	The session had lots of great tips to keep learners engaged, and also managing an online classroom. I need to convert all of our training courses to an online format this month so this was a very useful session.	From now on, I'll be teaching online. This session has helped me think about how to adapt my courses to be as engaging and useful as possible for the learners.
Stats ambassadors covid-19 response call	0.5	The RSS wanted to put out an article explaining covid numbers and terminology. In this meeting we discussed and planned the article.	Putting this article together forced me to simplify the complexity of Covid and pandemic modelling models into a simple, useful article. Practising this skill with fellow Statistical Ambassadors was useful as they all have different ways of communicating.	It's important that I can take complex terminology and models and explain them in a simple way to clients and their stakeholders. Taking part in this session helped develop those communication skills.
RStudio webinar on remote resources	0.5	An RStudio webinar on the different resources available for online teaching.	Mine shared some excellent resources with us which I intend to try out in my online teaching. She also gave some good tips of converting boring code into interactive online tutorials. In particular it was great to hear Mine's experiences of using RStudio cloud.	As a provider of online training, it's important for me to be up-to-date with the latest tools for delivering online training, and techniques to keep sessions engaging. This session will improve the quality of online training I can provide.
The Human Side of Statistical Consulting	4.0	I wanted to reflect on how I can be a better consultant, especially within the field statistics and data. I studied this book, reflected on my own experience and practice and made notes.	It was beneficial to reflect on what makes a good consultant. I reflected on the minimal required knowledge and identified topics which I could improve on. I also compared the suggested scoping questions with ones that I tend to use whilst scoping.	Reflecting on my practice as a statistical consultant mostly got me thinking about how I can best set expectations at the start of a project, and work with the client to ensure that the project best meets their needs.
Shiny Modules	3.0	I learnt about Shiny Modules, a way using functions in your Shiny apps to make you code more modular. I worked through a tutorial, read many articles, and implemented a demo example.	I regularly develop Shiny applications, but didn't know what Shiny modules were or how to use them. I am working on a large application for a client which lots of repeated code so this seemed like a great time to learn. I'll use my knowledge of Shiny modules in all my Shiny app development going forward.	Using Shiny modules means that my code is more modular. This makes it much easier to read, de-bug and test. Clients will benefit from cleaner code and a better final product.
Newcastle Data Science Meetup with Jared Lander	0.5	The (online) Newcastle Data Science Meetup with Jared Lander. Jared talked about different statistical models and how to best visualise their components.	Jared gave some excellent examples of visualising models which gave me ideas. In particular I liked his use of the coef plot to compare multiple models. The coefficient distribution plots are also a nice way to visualise model uncertainty.	When I implement a statistical model for a client, it's important that I am able to communicate the components of the model to the client. Visualisation is a great tool for this, and in particular I think that coefplot is a great technique for comparing model components.
RSS Liverpool Florence Nightingale	2.0	I was invited to present at an RSS local group talk on Data Viz and Florence Nightingale. As part of my preparation, I researched a lot about data visualisation techniques, and in particular polar area charts.	During the development of my talk, I learnt how to create polar area diagrams in R. I also learnt how to use the flipbookR package. This allows users to build a visualisation in stages.	Now that I know how to create flipbook animations, I'll use them in teaching scenarios to help learners understand the different components of ggplot visualisations.
Learning new dplyr functions	1.0	This year there was a significant release of a very popular R package - dplyr. I spent some time reading and playing with the new functionality to ensure I'm up to date. Examples include the across() function and the slice_*() family	In particular learning about the across() function has improved the quality my code. This function allows me to apply the same transformation to many columns, minimising code duplication.	As an R expert, it's essential that I am up-to-date with important developments in the popular R packages.
Managing a junior data scientist	5.0	In work I've begun managing a junior data scientist. I'm responsible for planning his work scheduling, monitoring his performance, and supporting his development.	Managing John has helped me reflect on workload scheduling and priorities. We have weekly scheduling sessions, where we choose the most important tasks, and estimate the time required to complete them. I'm having to choose a management style, improve my listening skills and balance supporting him whilst also enabling independent growth.	The listening and support skills which I'm developing in my role as a manager will enable me to better manage the teams I work with on client projects.
RSS Panel on Statistical Consultancy	1.0	As part of the RSS 2020 Conference, I took part in a panel session about Statistical Consultancy. Preparing for this talk made me reflect on my experiences as a statistical consultant and consider my strengths and weaknesses. I also learnt a lot from listening to the other panelists.	As part of my reflection I identified a time where things went wrong. A lack of communication with the client led to a misunderstanding, which left the client feeling unheard and ignored. I then reflected on how I can best communicate with clients in future. I believe the key to this is setting clear expectations at the start.	In future, I'll be clearer with the way I communicate with clients. I'll set expectations clearly at the start about communication as well as delivery. This should help my clients feel confident in my ability to deliver on time.
Efficient R	3.0	I taught a course on efficient programming in R. In preparation for this, I learnt material on benchmarking, object growth and parallel programming. Main resources were the books Efficient R Programming & Advanced R.	I've always had a vague understanding of efficiencies in R programming, however this period of learning forced me to really understand the main concepts. I learnt how R handles objects in memory, the best tools for benchmarking code, and the options available to R users wanting to parallelise code.	Having a good understanding of programming efficiencies in R will allow me to provide code reviews for my clients, write efficient code for them, and provide advise if they want to consider parallelisation.
R CI book	10.0	I am co-authoring a book on R and continuous integration with Colin Gillespie. This process involves researching how best to use tools like Git, Travis and GitHub Actions.	Researching, writing and editing a book has taught me so much about git and continuous integration. I've had to delve deep into topics in order to understand them well enough to teach them.	My knowledge of continuous integration and in particular git has dramatically improved. I have a much deeper understanding of git, which will help me set up sensible workflows for my clients.
Asynchronous Plumber	2.0	A colleague asked me to develop an asynchronous API for his project. As part of this I had to learn about asynchronous APIs, and the R packages {plumber} and {future}.	{plumber} is a package I've wanted to learn about for a while. It allows you to create web APIs directly from R. This increases my options when choosing a deployment method.	Clients often request deployment via API. I now have an understanding of how to deploy an API, and how to manage asynchronous API requests.
NHS R Conference	3.0	At the 2020 NHS R conference, I presented a workshop on ggplot2, and watched a number of sessions, include Silva's workshop on Xaringan.	I particularly benefited from Silvia Canelon's workshop on Xaringan. I use Xaringan often for presentation slides and am pretty up to date in terms of the relevant technical skills. However, her workshop introduced me to the design elements, and making my slides stand out. I've since created a set of slides in this more creative, stylish way and I feel it makes the content more engaging for listeners. There were also useful talks on data science (Lisa Clark) Qlik Sense (Jamie-Leigh Chapman) and R Code Quality (Colin Gillespie)	Many of my clients (both training and consultancy) work in Public Health. The NHS R Conference is massively useful for me to hear the challenges, progress and approaches arising in the public health sector. Hearing that it's hard for analysts to deploy shiny apps within the NHS encourages me to provide alternative solutions. Hearing managers worrying about security reminds me to support them with RStudio Package Manager. Although I do learn technical tips at this conference, the real benefit is understanding the users.
R Weekly Newsletter	6.0	I subscribe to the R Weekly newsletter. Once a fortnight, I sit down and review the recent articles, reading blogs and making notes on new information. Learning varies from package developments and new functions, to best practices for teaching.	The new R packages and functions I learn about in these sessions increase the size of my rstats toolbox.	As a leader in RStats consultancy and training, it's important for me to be up to date with the latest developments.
Mentoring with Benjamin	3.0	In 2020, I've been participating in peer mentoring sessions with Benjamin Wolfe who I met through #RStats. We meet once a month to check in on our data science projects, work, and self-development.	Although these started as an informal chat to stay in touch, I've found these sessions very beneficial to my professional development. We regularly reflect on working practice, statistical resources and data science best practice.	These reflection sessions have helped me consider which areas of my development to focus on, and how I can develop these skills to support my clients.
Kubernetes course	1.0	This session was run by an external company to teach Jumping Rivers how to use Kubernetes - a tool for managing virtual machines when deploying complex models.	This was a 4 hour session, however it was poorly pitched and delivered. I left with only a slightly better understanding of kubernetes then I already had. However, I did get a lot from reviewing the company's approach to online training and comparing this with our own approach. It was beneficial to participate in online training as an attendee rather than the presenter. It highlighted to me the importance of structure, setting expectations and regular breaks.	I learnt more about the basics of kubernetes so I could now confidently discuss requirements with clients. On a broader level, I also revisited our own online training sessions after this and made changes which should help make the sessions more engaging.
Mindfulness for Innovation	1.0	Mindfulness for Innovation was run by Alan Ross on behalf of the Innovation SuperNetwork. Alan led a number of meditations sessions, and we had a discussion about how mindfulness could support innovators.	I had some experience with mediation previously, but had stopped practising. I was mostly interested in attending this session to understand how meditation can help with innovation and deep work. However, it mostly highlighted to me that I am approaching burnout, and have a very restless mind. I've since re-introduced short meditation sessions into my working day. I've found that even short sessions can help prepare me for focussed, distraction free sessions.	Managing my mental health will me become a better, more focused and supportive consultant. I've also found that I'm able to better focus on hard tasks - working better.

Detailed Report: 2019

In 2019 I completed 100 hours of CPD training, across 5 categories.

	Activity Type	Learning Hours
	Professional activity	42.5
	Work based learning	19.5
	Formal / Educational	18.0
	Self-directed learning	15.0
	Other	5.0
Total	—	100.00

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
RSS Lancashire Chair	1.0	For 2019, I am the chair of the RSS Lancashire Local Group. This involved organising talks, marketing the talks and chairing meetings.	Running the RSS local committee involves organising events, chairing meetings and engaging with the local statistics community. All these things improve my organisational skills and broaden my professional network.	Volunteering on the local group committee showcases to my clients that I am an active member of the statistical community (especially in the North West). It also demonstrates my ability to be responsible and have good organisation skills. It's an elected position.
ISO9001	5.0	Digital Rail decided to apply for ISO9001 certification. I was put in charge of developing a quality management system and ensuring that it was robust enough to pass the certification. We passed first time and Digital Rail have ISO90012015 certification.	I hadn't written any quality assessment documents before so this was all new to me. I learnt how to write a quality management manual, and the key components required for one. I also learnt about the importance of constantly reviewing a QMS and the systems used to perform such reviews. Personally, I feel knowledge of QMSs and how to write and maintain them will be important in any future roles, especially consultancy.	This has huge benefit to users of my service. By writing a QMS, this showcases that providing a quality service is of great importance to me. Writing a QMS demonstrates how we will deliver a quality service, and what systems we have in place to ensure the service delivered meets the objectives.
Florence Nightingale	1.0	I presented a talk to a group of school children on Randomness and Probabilities.	Writing this talk made me think about the best way to explain complex concepts such as randomness to 16 year old girls. There were also a number of excellent talks which gave me inspirations for other styles of teaching I could adopt.	As a professional female mathematician I think it's important to demonstrate to young girls what sort of career paths are possible for them. I would hope that users of my service this see that as a worthwhile development, giving back to the STEM community. Finally it demonstrates to my clients that I am able to convey complex statistical concepts in an understandable, digestible way.
STOR-i Conference 2019	2.0	The Annual STOR-i conference held at Lancaster University. Approximately 100 attendees across academia and industry, mostly focused on cutting edge research in statistical and operations research. I attended a key-note and the poster session.	Even though I am now in industry, I believe it is important to stay up-to-date with the statistics literature. By attending this conference, I am able to keep up to date with the latest developments in change point analysis, extreme value theory, multi-armed bandits and cluster analysis.	Users can be assured that I keep up-to-date with the latest developments in statistics, and am well-versed in a number of statistical areas.
Supervising a placement student	3.0	Callum is a Computer Science student who is on placement at Digital Rail. It was my responsibility to supervise him and project manage his work.	I had to manage Callum's research project, ensuring he was on task, supporting both technically and emotionally. His project was written in a programming language I do not use (Java Script) so we have to ensure we were communicating well and often. Mentoring Callum helped me develop my supervision skills, my project planning skills and my java script knowledge!	Supervising Callum really helped me improve my project management skills. I had to set small achievable tasks for Callum, and monitor them frequently to ensure that the project was on track. Having strong project management skills is a key aspect of the service I provide to my clients as I am often the project lead.
UKRRIN Lecture Series - Centre of Excellence in Rolling Stock	2.0	UKRRIN organised a series of lectures at the Institute of Railway Research, Huddersfield University. Topics included risk modelling, novel braking systems and a wheel motor project.	As a data scientist, I frequently have to work on domain specific problems. In my current role, I specialise in railways, although I have no formal engineering or railway experience. These talks broadened my domain specific knowledge, this will help me work in this sector in future, and enable me to communicate with railway engineers.	As a digital railway data scientist, having good domain knowledge is key. By attending this lecture series, I demonstrate to users of my service that I am willing to develop my domain knowledge.
DSI talk Quantifying Input Modelling error in Simulation models	0.5	Lucy Morgan, a new lecturer in the Management School, gave a very interesting talk on simulation models. Lucy discussed how assessing inputs into a simulation model is an area often overlooked, both in terms of the uncertainty and biases that inputs have on model outputs. Lucy showed how these effects can be properly quantified and interpreted.	Input error is an important but often neglected area of simulation modelling. Learning about what input error is and how I can best model it will be key to any future simulation modelling that I perform.	Any clients requiring simulation modelling will be reassured that I have thought about the possibility of input error, and have taken steps to quantify their effects on the model output.
Manchester R User group	1.5	There was an hour long workshop on Shiny Reactivity before the talks. Chris Billingham (MAG-O) spoke about using tidytext on musical lyrics. Jane Jin Autotrader spoke about decision making multinomial logistic regression.	The Shiny tutorial by Chris was really useful. Interactivity is a key component in shiny development and wasn't something I had understood fully before.thought I thought I was Jane introduced the mlogit package for multinomial logistic regression which I hadn't used. Chris Billingham's talk gave me more of an insight into the capabilities of the tidytext package and also introduced me to the tuneR package for analysing music and speech. I also had a good networking session with Chris and Jonathan from MAG-O about data pipelines.	Now I have a better understanding of how Shiny interactivity works, I'll be able to build faster, more robust apps for my clients.
RSS Lancashire Talk Peter Van der Heijden	0.5	Peter gave a talk about population size estimation where linking registers results in incomplete covariates.	Peter's talk introduced me to a new challenge in population estimation, and how linking multiple registers and using the Expectation Maximisation algorithm can be used to solve this. These methods are applied to hard-to-reach populations. Applications include victims of human trafficking, the homeless, domestic violence, illegal immigrants etc. It was interesting to think about this problem from a modelling perspective.	If users of my service came to me with a population estimations problem for a hard-to-reach population, e.g shelter estimating homelessness - I'd know what research to consider, and how to decide if the model was appropriate.
DSI talk Data Science and the ACM	0.5	Vicki Hanson spoke on the ACM’s Data Science initiatives at the DSI morning talks.	Vicki talked mostly about the challenges when setting up a Data Science masters programme from the perspective of a computer scientist. This encouraged me to reflect on my own data science learning journey, one pre-dominantly from the statistics side. It also made me consider the best way to teach these concepts when time is short.	Thinking about the best way to teach data science will structure and inform the way that I teach both statistics and programming in my training courses.
Data Conversation Best Laid Plans	1.0	The session took the form of structured discussions and activities rather than talks. Topics included preparing data for collection, identifying existing suitable datasets, and ensuring that our data is well managed through the whole lifecycle.	Managing data sets is a key role of a statistician. The discussions I had with my peers helped me reflect on the data sets I currently depend on, and whether I'm storing them in the optimal way to support analysis both known and unknown.	Often when supporting clients with statistical consultancy, they haven't identified their data source . This session helped me think abut how I can best ensure that the correct data is collected first time around, minimising effort and cost for the client.
DSI Company Partners' Workshop	1.5	An afternoon of talks and activities focusing on the future of Applied Data Science. Talk topics varied from explaining existing and solved problems in industry, to incorporating data science into future projects. There was a workshop activity called "Building and Implementing Data Science Solutions" where we were given a data science problem and had to build the right resources in terms of both skilled workers and infrastructure.	I found the workshop activity very enlightening and enjoyable. The process of building a team that could provide a suitable level of support was challenging and informative. This made me reflect on which projects require highly skilled statisticians vs junior data scientists and the importance of a good project manager. The talks introduced me a broad range of statistical challenges in industry.	Being able to select the best tools for the job in terms of both human resources and infrastructure is key to running a successful consultancy project.
RSS Lancashire Talk Claire Miller	0.5	Claire spoke about global lake water quality and the statistical challenges associated with modelling ecological systems. The GloboLakes Project used long-term satellite observations in order to estimate the water quality of lakes and look for spatial similarities. Claire showed us visualisations of both the satellite data and the water samples over time, and highlighted the difficulties of matching up the insitu samples with remote sensing data.	Claire discussed the many challenges faced including dimensionality and missing data. For example, one of her sensors stopped working early on in the project, forcing her to find an alternative data source. This reminded me of the importance of planning ahead and making project risk assessments, to ensure that when challenges occur, there is a potential back up. Claire also discussed how making her research accessible via the development of an R package impacted her project. This reminded me the importance of open-source work and encouraged me to consider developing packages where appropriate.	Creating a risk assessment at the start of a consultancy project will reassure clients that I have considered the potential challenges and am prepared to handle any unexpected problems arising.
Challenges in practical applied data science with Philip Jonathan	0.5	Phil talked about how the digitalisation agenda and data science is impacting organisations like Shell. He summarised the areas where data science is proving most beneficial, and the current difficulties	Phil's discussion about methods for model assurance was very interesting. It encouraged me to reflect on the importance of algorithm agnostic assessment.	When building a model for a client, validation and assessment is key. This talk presented a number of methods to consider when comparing data-driven solutions.
DSI talk on predicting debt repayment	0.5	A talk from Edmund Ryan on predicting future repayment patterns of portfolios of non-performing debt.	I found it interesting that Edmund's statistical model considered fairness as well as profitability, setting a minimum repayment of £1 per month. This made my reflect on the importance of adding ethical and human components into statistical models. There was a good discussion about methods for speeding up stochastic statistical models.	Knowing how to implement stochastical models in an efficient manner could be useful on client projects. Also an awareness of keeping models fair and ethical is very important.
SatRday Newcastle	3.0	A one-day conference on the R language, held in Newcastle. Key talks included Geocomputation Robin Lovelace, Shiny deployment with Seb and creative coding from	Thomas's encouraged me to think about the way I approach projects. His talk inspired me to try more creative R projects and see if I can reap the learning benefits associated with them.them Seb's talk was very useful practically to me as I need to understand different deployment strategies for Shiny apps in my new role.	Robins's talk introduced me to state of the art methods for handling and manipulating spatial data. Many of my clients want to wrangle and visualise spatial data so my increased knowledge will allow me to support this.
Parsnip and Recipes	3.0	Self-taught research to learn the {recipes} and {parsnip} packages which make up the tidymodels framework in R. I read vignettes, blog posts and worked through practicals.	I had used the {caret} package previously but did not know the new tidymodels framework. I'll need to teach these packages at Jumping Rivers, so I need to be very comfortable using these packages.	Our clients are often interested in learning how to implement machine learning models in R. This framework allows them to compare different models in a comparable space. I'll also be able to help our users convert over from the older package {caret} to the new framework {parsnip}.
VM training with Seb	1.0	Training on how to set up and host a Virtual Machine for training courses.	As part of my role at JR, I'll be supporting Seb with some of the infrastructure for our training courses. Seb explained the full process of developing VMs which includes building a virtual machine, setting a DNS, adding SSL certificates and installing dependencies onto the machine.	Stable VMs are essential for a smooth training experience during courses. It also means I can have the technical skills to fix any issues that may occur during training. Finally, I'll be able to discuss deployment elements with consultancy clients in a more technical manner.
RStudio Instructor Training	5.0	A seven hour course delivered by Greg Wilson on how to become an RStudio accredited instructor.	This course introduced me to many techniques useful for teaching tech. New ideas included learning personas, mental models and formative assessment. I've since incorporated all of these into the courses that I deliver at JR.	Having an understanding the different ways that people learn, and how best to assess their understanding should help me deliver more engaging, successful training courses.
Organiser of R-Ladies Lancaster	8.0	I'm currently the lead organiser of the RLadies Lancaster meetup. We ran our launch event in May, and in 2019 we ran six events. We've led practical workshops on {dplyr}, {ggplot2}, {tidyr}, {stringr} and {lubridate}.	As a professional female programmer, I believe it's important to give back to the community. By running these sessions, I have to design educational content at a suitable level, and advertise the events. R-Ladies is also a great place way to get to know the local R community and widen my professional network. I also learn lots each session, this could be a new R package that I haven't heard of, or simply a better way to explain a problem to an attendee.	Organising a free, open group allows members of the public to get support with R. By delivering workshops to the community and supporting new learners, my clients can see that open-source development is important to me, that I'm a skilled educator and that I'm a supportive listener. It might be important to clients that I support STEM education programmes for women. This also demonstrates to clients that I an active member in the R North West community.
KableExtra	1.0	A project for client meant I had to learn more about the R package kableExtra. I read the package vignettes, investigated the issue in the package and spent time testing the different options available.	There are many ways to display tables in R Markdown documents, but one of the most popular is {kableExtra}. Having a deep understanding of all the possible styling options, allows me to branded, complex tables in both HTML and PDF.	I'm able to support users with technical questions related to formatting of both HTML and PDF tables using {kableExtra}. I'll be able to help clients meet specific table styling requirements programmatically.
RSS Lancashire Machine Learning in Astronomy	0.5	The RSS Statistical Computing Section and the RSS Lancashire and Cumbria local group jointly ran a workshop on Machine Learning in Astronomy. The aim of this event was to bring together statisticians, machine learners and physicists to introduce academics and PhD students to the range of research challenges that sit at the interface of these research fields. I was one of the organisers.	It was really interesting to hear about the statistical challenges being met by physicists working in astronomy. I was particularly interested in the deep learning image classification tool PlanetNet by Ingo Waldmann. This made me more aware of the how statistical techniques can be adapted to the physics domain.	Having an understanding of some of the statistical challenges arising in Astronomy, and the existing work being done will allow me to support any clients working in that field.
Tidyverse Exam	0.5	This was a formal exam to assess my ability to code using the Tidyverse R libraries. I was assessed by Dr. Greg Wilson and passed first time.	Having a formal qualification in the Tidyverse libraries is a great achievement for me. It documents my ability to use the tidyverse.	Completing this exam has enabled me to become a Certified RStudio Tidyverse Trainer. Having this qualification will confirm to clients that I am good at teaching programming and that I have a good understanding of the R tidyverse packages. This will give them confidence when hiring me to deliver training that I am skilled at what I do.
Introduction to Python	8.0	I attended a two day Introduction to Python course run by Jack Walton. I have already done a little python programming but had not received formal training before.	As a data scientist, I frequently have to edit, format and write Python code. Having formal training means I have more confidence in my coding and have a better understanding of the language.	Clients frequently come to us with Python code. By having a more formal knowledge of the language I will be able adapt their code and integrate it wit our data pipelines more easily.
OR Analytics Summit	1.0	The annual one-day Analytics Summit organised by the OR Society. I attended talks and ran an exhibit.	Payal Jain gave an interesting talk defining the building blocks for an effective data analytics function. She highlighted the importance of focusing on people rather than data, which made me reflect on my own approach to large scale data science projects.	The OR community contains a number of clients who are keen to migrate from other software and learn R & Python. By engaging with this community and learning about the challenges in the industry, I can better support my clients needs.
RStudio teaching exam	1.0	This was a formal exam to assess my ability to teach R programming.I was assessed by Dr. Greg Wilson and passed first time.	In order to prepare for this teaching exam, I had to prepare a lesson, draw a concept map, design interactive questions and create a learning persona. I found this process really beneficial to thinking about the way I develop learning resources.	Completing this exam has enabled me to become a Certified RStudio Tidyverse Trainer. Having this qualification will confirm to clients that I am good at teaching programming. This will give them confidence when hiring me to deliver training that I am skilled at what I do.
Observing training	4.0	I observed my colleague Theo deliver training for two days at Santander. The courses were Introduction to R and Programming with R.	In my new position, I'll need to deliver training courses. Observing Theo, I got to witness his delivery style, become more familiar with our training materials and understand the flow of an on-site training course. I also taught one chapter of the course which was great for my confidence.	Observing others deliver training, and learning the teaching material will help me deliver excellent quality training courses to our clients.
Unix basics	2.0	I worked through the online materials "Shell Novice" put together by the Data Carpentries.	I often have to work in the Unix shell whilst using git or navigating around my laptop. I wanted to increase my knowledge of the shell to allow me to work more efficiently, and learn more about writing shell scripts.	Having good shell scripting skills will allow me to work more efficiently on client projects, and write bash scripts as part of consultancy projects.
Learning the {purrr} package	3.0	I needed to learn the {purrr} R package for work, both for client projects and teaching. I worked through Chapters 21 and 25 of the R4DS book.	The {purrr} package is a key part of the tidyverse and allows users to quickly write iterative code. Learning {purrr} will allow me to quickly write iterative code, in the new tidyverse framework. It's also a package that I'll need to teach in one of our training courses.	Being able to use {purrr} will allow me to re-write slow for loops in a more efficient, tidyverse framework on client projects. I can also write code to help users complete common challenges, such as combining similar datasets.
useR! 2019	8.0	I presented a workshop on Travis CI at useR! 2019 held in Toulouse. There were many fantastic, relevant talks and I left this conference with lots of new learnings.learnings Useful talks included combining Shiny with R Markdown, new developments in {dplyr}, enhancements to {tidyr}, {golem}, {goodpractise} for package development and reusing tidyverse code with {rlang}. code	I left the conference with a huge list of R packages to investigate further. The talks on best practise gave me excellent advise which I have since incorporated into my coding.	As an expert in R development, it's essential for me to be informed on the latest developments. Attending the useR! conference is one of the best ways for me to update my knowledge and engage with the R community.
Shiny exam	0.5	This was a formal exam to assess my ability to code using R Shiny. I was assessed by Dr. Greg Wilson and passed first time.	Having a formal qualification in the Shiny is a great achievement for me. It formally documents my ability to use Shiny.	Completing this exam has enabled me to become a Certified RStudio Tidyverse Trainer. Having this qualification will confirm to clients that I am good at teaching programming and that I have a good understanding of R Shiny. This will give them confidence when hiring me to deliver training that I am skilled at what I do.
Learning the {broom} package	1.0	The {broom} R package is used to manipulate the output of statistical tests in R. I learned how to pull out statistical parameters such as the AIC and BIC. I also learned how to calculate the model residuals and compare them to the original data set.	The {broom} package is very helpful for quickly comparing the output of statistical models. Understanding how this package works will help me create useful diagnostic plots and analyse model fit.	This package helps users interpret model output in a more user-friendly way. I can know better support users in working with the standardised residuals of their models. This is also part of the curriculum on our Statistical Modelling course.
Spatial Data Learning	5.0	In work I had to write material for a new course on Spatial Data Analysis. This involved a lot of research into the best ways to handle spatial data in R. I used a number of resources, but my main focus was reading the Geocomputation with R book.	Whilst writing this course I learnt about a number of R packages for handling spatial data, in particular {sf}, {tmap} and {leaflet}. I also learnt about the different ways of coding spatial data (vector vs raster) and how to use co-ordinate reference systems.	Many of our clients with to build maps to visualise spatial data. I now have a good understand of creating both static and interactive maps in R. I've also got good knowledge around how spatial vector data is handled in R, and how to troubleshoot common problems.
Royal Statistical Society Annual Conference	5.0	I attended the annual RSS conference, this year held in Belfast. The panel discussion "stats the way to write it" was a fascinating discuss of how to best write a book on a statistical topic. I also particularly enjoyed the session "Show me the stats" which focussed on visualisation. This included talks from Allan Reese and Robert Cuffe(BBC). Finally, "Fighting fake news and false facts with evidence and statistical thinking" was a great discussion about the challenges of communicating statistics in the media.	The annual RSS conference is an excellent opportunity for me to learn about new statistical techniques across various fields. It also helps me develop my teaching and communication skills, and is an excellent networking event.	Improving my statistical communication skills will help me better work with clients on technical projects. Also, keeping up-to-date with the newest statistical techniques will widen the tools I'm able to apply in consultancy.
EARL Conference	2.0	Annual conference on the Enterprise Applications of the R Language, hosted by Mango. I was exhibiting a stall and attended a number of talks. The talks I found most useful were Julia Silge discussing the Stack Overflow Developer Survey and Kelly O’ Briant covering how CI/CD tools can enhance reproducibility for R and data science.	Kelly's talk discussed best practices around getting data science projects into production. I learned about strategies such as green-blue deployment and methods for continuous deployment. It was also an excellent networking opportunity.	Having a good understanding of the best practices and possible pitfalls when deploying data science projects is key to ensuring a successful production pipeline for clients.
Learn Lintr	1.0	In order to improve the R packages we write at Jumping Rivers, I had to learn about the {lintr} package, in particular, the update to 2.0.0.	Code style is a key component of developing good quality code. By using the {lintr} package, I'm able to force all code that I create to follow these code conventions.	Writing code which follows a strict coding style will help ensure that code I write for clients is clean and readable. Following code conventions when delivering training also showcases best practise to course attendees.
NHS-R Conference 2019	2.0	I delivered a workshop on {ggplot2} at the two-day NHS R conference held in Birmingham. I also attended a number of interesting talks including one on Shiny by Nic Crane and one on Why R by Steph Locke.	I learnt some good tips and tricks for Shiny development from Nic's talk. I also heard first hand the challenges that NHS analysts meet when trying to develop using R. For example, no access to modern packages, no way to deploy shiny apps, I.T infrastructure issues.	Many of my clients work in the public health sector. Having a good understanding of the challenges they face daily help me empathise with them, and offer appropriate solutions.
RStudio Team System Administration Course	3.0	As part of my role supporting Seb and the data engineering team, I took the RStudio Team system administration course. I learned how to administer RStudio products such as Server Pro, Connect and Package Manager.	It was really useful for me to learn how these products which I use daily are administered. I learned more about authentication methods such as LDAP and security best practices.	Many of our clients use the RStudio Team products. Having a better understanding of the infrastructure behind these products and the different license options will enable me to support clients using these tools.
RStudio Technical Exchange	1.0	In 2019, I attended three of the RStudio technical exchange meetings. In these sessions, staff at RStudio introduce new developments coming to the RStudio ecosystem. Discussions have included reticulate, installing binary R packages with RStudio Package Manager and how to pin data on RStudio Connect.	The technical exchange keeps me up-to-date with new developments in the RStudio ecosystem. I've learned how to use pins to pin data on RStudio connect, and the new developments to the {plumber} package for developing APIs.	Many of our clients use RStudio products daily. Having a good understanding of their features, and the upcoming developments allows me to support my clients needs. This session also allows me to feedback any feature requests from our clients directly to RStudio.
Career Mentor Launch	1.0	Introduction for the career mentoring scheme I'm taking part in. There was a one-hour session led by Rob Pickersgill There was also networking time with other career mentors which I found really beneficial.	Rob's session on mentoring was really helpful. He discussed techniques such as empathy and active listening. He also encouraged me to reflect on what I wanted to get out of the scheme.	I believe there is a strong link between mentoring and being a good consultant. Rob's talk, and chatting with the other mentors gave me time to reflect on my consultancy skills, and in particular active listening.
Career Mentoring	2.0	Career Mentoring. I've been mentoring two students who are interested in becoming data scientists. One is an undergraduate, and one is a post-graduate. The scheme ran from November 2019 - May 2020. After completing his PhD, Ben was successful in securing a data science job.	I've had to practise my active listening skills and ensure that I support the students without pushing my own thoughts onto them. I've practised my editing skills, helping both Ben and Theo with their C.Vs and cover letters. I developed my project management skills, supporting Ben in the process of selecting and applying for a job.	Being a career mentor showcases that I'm keen to support junior data scientists in their career. It also demonstrates that I'm supportive and a good listener.
SatRday Cardiff	3.0	I presented a talk at the one-day conference on R held in Cardiff. Useful talks included Dominik Krzemiński's talk on debugging Shiny applications and Arun Srinivasan's talk on the {datatable} package. Chris Jackson also gave a though provoking talk on impostor syndrome.	Arun gave a thorough introduction to {datatable}. It's not a package I've used before, as it's an alternative to the tidyverse framework. It was helpful to learn about the speed up in manipulation it can provide, and seen the (very different) approach to the code style. In work I do a lot of shiny applications so Dom's talk on debugging was very helpful.	Being able to debug shiny apps efficiently will improve the development process.
Manchester R November 2019	1.0	I delivered a talk on Continuous Integration. Chris Campbell gave a talk on R for Primary Teaching Phrances Ayvee Perez talked about reactive function in {shiny} and tips for best practise	It was brilliant to hear Chris talk about how R and {shiny} could be used to help KS1 children learn mathematics. His creative {shiny} apps have given me ideas about how we could be more creative in the way we teach. I also had some useful networking discussions on the CI/CD approach I talked about.	By constantly reflecting on how we use technology to teach statistical principles, it allows me to be more creative with my teaching and hopefully deliver high quality training sessions.
PyData Lancaster	0.5	David Elliott gave a fascinating talk on Seizure Detection with Machine Learning. Ryan Callihan gave an excellent overview of Natural Language Processing.	I found it really useful to hear Ryan discuss the different methods he uses in NLP and particularly the different libraries that are used in Python. It was also an excellent way to meet other data practitioners in Lancaster.	Many of our clients work with text data. I now have a better understanding of the different models available and how they should be applied.
Why R	1.0	I shadowed Colin Gillespie delivering a Why R course. This is a course for managers explaining the positives and challenges when bringing R into their teams.	This is a course I need to deliver next year, so it was great to observe Colin's training style. It gave me ideas of how to communicate and connect with managers and it highlighted the concerns that managers often have about adopting a programming language into their analytics teams.	Managers often have concerns when adopting a new product into their team. Being able to provide an informative WhyR session will help them be able to make an informed decision about this choice.
SQL murder mystery	1.0	I decided I wanted to brush up on SQL. I completed the SQL murder mystery tutorial online.	I haven't used SQL before. I wanted to gain basic knowledge of the commands, so I can communicate better with clients and attendees at my R training.	Many of our clients use SQL, so having a basic understanding of SQL commands can help me understand their data bases. Also, when teaching people how to use R, many of them know SQL, but not R. By understanding basic SQL, I can make comparisons between R and SQL, hopefully cementing their knowledge. This is particularly relevant when teaching the dplyr & tidyr package.

Detailed Report: 2018

In 2018 I completed 61.5 hours of CPD training, across 5 categories.

	Activity Type	Learning Hours
	Professional activity	39.0
	Formal / Educational	16.0
	Work based learning	5.0
	Other	1.0
	Self-directed learning	0.5
Total	—	61.50

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
STOR-i CDT Conference	1.0	Poster session for the CDT STOR-i Conference	I discussed a number of new techniques being developed in the STOR-i CDT, including a multi-armed bandits method and new clustering techniques. There is a possibility to set up a PhD project or internship with a STOR-i PhD student.	Close collaborating with Lancaster University allows us to incorporate the latest academic algorithms into our work.
R Ladies Manchester January	1.0	Using a real-life data set R-Ladies Manchester hosted a workshop evening allowing members and develop for data science. This session focused on using branches in GitHub to work on different versions of the project & Importing data into R.	The introduction to using branches in GitHub was very insightful. I also taught some data analysis techniques to other attendees.	Using GitHub is an excellent way to ensure that our service is well documented, uses version control and is reproducible.
STEM Ambassador Careers Fair	0.5	Careers event for Years 9-11. I volunteered to give a 20 minutes presentation about my career.	As part of delivering this talk, I had to work on my communication skills. I decided to present without slides, and therefore it was critical to ensure I was able to communicate the pathway that statistics can take you clearly to the students.I also believe that it is greatly important to give back to the STEM community, and particular engaging with young STEM students.	Better communication skills, and it's good to give back.
RSS PSF Route to Chartered CStat	2.5	I watched videos, read articles on the RSS website and did some personal reflection about my CPD.	I learnt how to track my CPD, what sort of activities are applicable, and how more about the pathway to CStat. I began to re-think how I present myself as a professional statistician, and the guidelines to follow as a Fellow of the Royal Statistical Society.	CPD is vital to provide a professional, up-to-date, service to clients.
Her + Data MCR	1.0	Their were talks from Silvia Maggi (computational neuroscientist), Rebecca Davey (data scientist at INRIX) and Clara Higuera Cabanes & Magda Piatkowska from the BBC.	Reconnecting with Magda and Clara was great. Their talk about clustering, really got me thinking about using data to personalise services. Rebecca's talk was particularly relevant as it was in the transport industry. Silvia's talk drew some fascinating similarities between neural networks and the brain.	These talks gave me encouraged me to start working with neural networks, a new technique which we can incorporate in our work.
RSS PSF Effective Communication	0.5	An RSS sessions on effective communication. Speakers included Robert Cuffe (BBC News), Liberty Vittert (University of Glasgow) and David Beckett (Office for National Statistics). This webinar demonstrated case studies from the RSS Statistical ambassador scheme and highlighted how communication is a very important aspect of everyone’s careers.	These talks gave some excellent examples of how best to communicate statistical concepts.	Communicating is key skill when working with clients. It's vital that I can communicate the assumptions and reasons for choosing a model, as well as explaining the results and choices.
R4DS tutoring	1.0	I volunteered to tutor online for 2 hours as an R expert	The tutoring session involved me re-visiting the Tidyverse. Explaining these concepts to a beginner helped my improve my communication and tutoring skills. Explaining programming concepts over a messaging service can be challenging, and practising this allowed me to trial new techniques.	Improved communication, especially relating to coding and programming is vital for delivering an understandable service.
Manchester R May	3.0	Talks on RStudio Projects (Rachel Naylor), Interpretable Machine Learning (Pantelis Hadjipantelis) and Storytelling with R (Rob England). There was also a tutorial on sparklyr by Chris Campbell.	The sparklyr workshop was an excellent introduction to using Spark with in RStudio. I aim to add this to my toolbox of techniques. The RStudio Projects talk from Rachel gave some great tips on how to improve my quality assurance.	Spark is a big data technology. Using it will allow us to handle large data sets in a much more computationally efficient way.
RSS PSF Excellence in Official Statistics	0.5	Talks from winners of the Excellence in Official Statistics 2017 Awards. They discussed motivation for their work, how it was done, the key outcomes and how things have developed since.	Introduced a couple of new visualisation ideas that I hadn't seen before including heatmapping of SIMD.	New visualisation techniques can allow us to communicate our findings to clients.
Manc ML May	2.0	Using ML for image recognition to improve Google shopping descriptions (DreamAgility), Deep Feature Synthesis (Martin Eastwood) and GANS (Erik from Octavia)	Martin gave an overview of Max Kanter and Kalyan Veeramachaneni's paper on Deep Feature Synthesis. This is a new technique which I am keen to try out on our data. Erik gave a great introduction to GANS (Generative adversarial networks). This again was a new technique which I was hoping to learn more about. Erik's talk gave a good overview of the pros, cons, and when these techniques are appropriate. The Dream agility talk gave a great end to end story of how A/B testing can work. All three talks introduced new techniques which I am keen to introduce into our work.	Automatic feature extraction will allow us to generate better quality and more creative features. This will allow us to extract more information from client data.
R-Ladies	0.5	The session covered topics on Importing Data and working with Tibbles (chapters 7 & 8 of the R4DS book).	This book club session included an excellent comparison of data.frames() and tibbles(). This understanding and the extra tips will help improve my R coding.	Staying up-to-date with R coding practice is essential to good development.
RSSB Forecasting Workshop	5.0	This was a workshop on the forecasting adhesion on the rails, a key issue in the railway industry. The morning consisted of talks explaining the problem in more detail. This included an overview of operational problems (John Edgley, chair of AWG), the impact on train operators (Rob Cummings, Northern Rail), developments on the tube (Priyank Patel, TfL) and the Met Office Approach (Dr Victoria Chapman & Dan Adamson, Met Office). In the afternoon we took part in workshops to try and brainstorm potential approaches to the forecasting problem.	This session involved a lot of domain-specific knowledge. I left the event feeling like I had a much greater understanding of the rail sector and the many different players. The session on the MetDesk showcased a great way to provide domain-specific predictions in the rail sector. I learnt a great deal about the adhesion problem, and the RSSB funding available.	This event informed me on how we can best submit a funding call to forecast poor adhesion, If funded, we will be able to deliver a new product for our users, increasing safety and decreasing train delays during the difficult Autumn season.
UCREL Summer School in corpus-based NLP	16.0	A series of talks and tutorials on corpus-based NLP. Included sessions on social media analysis, authorship analysis for online text, semantic tagging, multilinguality, sentiment analysis and machine learning.	The tutorials were very informative and taught me how to implement a variety of NLP techniques in javascript. I have learnt new techniques in NLP and improved my competency in javascript coding.	A product that we develop is based on Natural Language Processing. The training that I received over the 4 days introduced me to new techniques that we can incorporate into our products.
Digital Rail Summit	5.0	An all-day, company wide summit. Sessions included SWOT analysis, value propositions, routes to market and how to make a business valuable.	As a new employee to Digital Rail, this was an invaluable experience to help me understand the full business, the different areas and challenges.	I now have a more thorough understanding of the market that Digital Rail operates in. In particular, the session on value propositions was really insightful. We have begun to build this into our product development, to ensure that we are meeting the customers' needs at all time by matching customer needs with a product solution.
UKRRIN Rail Data Challenge Workshop	1.0	The Centre of Excellence in Digital Systems (CEDS) at BCRRE hosted a workshop on the Rail Data Challenge. This challenge is an open invitation for innovators, developers and tech experts to develop new technologies to address three current rail challenges. The workshop introduced the challenges, available data sets, potential partners (DfT, RDG, RSSB) the available data and facilitated breakout sessions.	This session taught me about UKKRIN, a university research group in rail. Learning about the different projects they were working on is useful as we may need to team up with some of their academics in future projects. We also had a break out session where we discussed the problems in the rail industry and brainstormed technologies which would be useful in solving them.	Working with large research organisations expands the type of products which we can develop. Developing new technologies means we can offer new products to our clients which meet the needs of the rail industry.
OR60 Conference	2.0	I attended part of the Operational Research 60th Conference. Particular talks of interest were from TfM, Edwin, and the Women in OR panel.	Edwin is researching re-scheduling in the train delay scenario. It was useful to see how the problem is formulated in linear algebra. The Women in OR meeting was beneficial to my confidence and networking.	Staying up-to-date with the latest academic research allows me to bring bespoke techniques to our clients.
Data Conversations Keep it, throw it, put it in the vault.	1.0	This session focussed on the long term value of data. Questions discussed included how to make sure data is accessible in the future, should we keep all our data, and how can we assess what has long term value?	The speakers discussed their experiences with data, particularly how storing data that they hadn't expected to be useful was important in their analysis. We had some interesting discussions about lost data and personal data.	One of the first steps with users is discussing what data they have or deciding what data should be collected. Discussing best practices of data storage and collection will ensure we make the most of all available client data.
RSS Lancashire Talk Davide Pigoli	0.5	Davide Pigoli spoke about analysing of acoustic phonetic data. He was exploring the phonetic changes between different languages using audio recordings. Davide's approach used the log-spectrograms of speech recordings and compared the frequency covariance functions.	It was interesting to learn how it is possible to analyse acoustical data using Fourier analysis. In particular, separating word differences and language differences.	Expanding the types of data that we are able to analysis broadens the services that we are able to provide.
First of a Kind (FoaK) 2 Rail Competition Briefing - York	2.0	A briefing day for the First of a Kind (FoaK) 2 Competition. The competition was to identifying high TRL products and test them in real-world railway environment. The session explained more detail about the competition, how to apply and had a session on grant writing.	This session taught me the essentials of writing a government grant, skills of which I require regularly in my current role.	Writing good government grants (and hopefully getting them!) showcases that we know how to successfully take a project from idea to completion.
IMechE NW RD - Challenging Times	2.0	I was invited to present at the NW RD of the IMechE. I presented a case study on a vision system. There were two other case studies, which were followed by the Chairman's address. The Railway Division Chair, Andy Mellors, examined the economic and wider societal benefits which the rail industry delivers and considered the current challenges facing the industry.	The chairman's address by Andy Mellors gave me an excellent insight into the challenges that working in the railway can bring. I met a PhD student who presented his research on using deep neural networks to identify bicycles. He introduced me to a couple of techniques which may be useful in our work.	Presenting our research to members of the IMechE allowed me to understand better the difficulties in getting our research on the trains.
Code for Lancaster	1.0	Code for Lancaster is a volunteer Civic Technology meetup. This was the first session, and discussions revolved around “How might we better enable locals to be aware of local events”.	Useful networking session, mixing with local full stack developers as well as data scientists and statisticians.	A number of the topics came back to discussing the What services can we provide vs What services do the users actually want? These discussions made me re-think the services that we currently provide and encouraged me to focus on the customer.
Northwest Universities R Day	7.0	There were keynotes from Andy Field and Robin Lovelace, as well as flash talks, a poster session and workshops in the afternoon. I attended Chris Campell's workshop on The curse of the unspecified data definition.	This was a hugely beneficial day for me. The talks covered a broad range of topics and techniques from interactive teaching with the R language, to reproducible data analysis. I have plenty of great resources to take away and work through. It was also a great networking opportunity.	In particular the workshop with Chris Campbell taught me the importance of clearly defined data rules when dealing with client data. By implementing the principles and techniques that I learnt in the workshop, we will be able to provide a more robust data service for our clients.
STEM Ambassador Engineering Your Future	1.0	I volunteered at Engineering Your Future Career Awareness Days, one at Siemends and one at MMU. At each session their were 100 16-18 year old students all studying STEM subjects. I worked on the career stall and engaged with students about the different pathways that engineering could take them.	I learnt how to build a quick and simple object detection algorithm which was useful as we are intending to build or own bespoke object detection algorithm. I also practised my engagement skills, chatting with 16-year-olds about pursuing a career in engineering.	Being able to explain our work to non-technical users is important for engaging with clients. It's also important to show our users that we meet Corporate Social Responsible and give back to the future STEM students.
Rail Supplier Engagement Event	2.0	The Centre for Innovation in Rail at the University of Huddersfield ran an event for Rail Suppliers and SMEs. The session included a tour of the research facilities, gave an in depth overview of key industry partners and highlighted the research project currently happening at the Centre for Innovation.Talk topics included product testing, funding opportunities, routes to market and product support available from Rail Alliance.	Martin Little from the Rail Alliance gave a thorough introduction to the rail industry. He explained TOCS, ROCS and government agencies. He also gave 15 tips for getting started working in the rail industry. This overview really helped me understand how the many different rail companies interact.	A detailed understanding of the interaction between companies is vital to ensuring that we are working in the correct space and all the appropriate partners. I also spoke to Rail Alliance about how we can get testing vouchers to test our products. By testing our products on a real train we can improve the quality and highlight any potential issues early.
RSS Stats Ambassador Stat of the Year	2.0	Assisted the RSS in checking their statistics for Stat of the Year 2018.	This activity allowed me to develop my fact-checking skills, quickly looking for potential errors or pitfalls in the statistics presented. Due to time constraints, this also forced me to practise working efficient and develop my time-keeping skills.	Having a keen eye for possible errors or biases in data is an important skill for ensuring that the analysis and statistics that you have developed are a fair representation of the data. It was great to practice ensuring that the answer provided actually answers the original question! This is another vital service when providing consultancy for clients.
RSS CPD training webcast	0.5	I independently worked through CPD training on the RSS YouTube channel.	Although I have already attended an RSS CPD course, I found it very useful to revisit this webcast. This time I focussed more on the planning and general development rather than the actual activities.	Since working through this I have decided to make a development plan in order to ensure that my CPD matched the skills that I need to develop.

Detailed Report: 2017

In 2017 I completed 62.5 hours of CPD training, across 5 categories.

	Activity Type	Learning Hours
	Work based learning	23.0
	Self-directed learning	14.0
	Professional activity	12.5
	Other	8.0
	Formal / Educational	5.0
Total	—	62.50

Benefits Gained report

Title	Learning Hours	Activity Description	Benefit to Practice	Benefit to Users
UniBrass Trustee	5.0	I served as secretary and trustee for the UniBrass Foundation. The UniBrass Foundation set up to advance, improve, develop and maintain public education in, and appreciation of, the art of the brass band movement, especially at universities. They aim to do this primarily by organising the annual UniBrass Championships, offering support to existing university ensembles, and encouraging the creation of new university bands.	As secretary, I was responsible for writing minutes, liaising with the Charity Commission and ensuring that the constitution is up to date.	As part of my role, I had to learn how to write GDPR and child protection policies. I also had to learn how to update the charitable objectives. These writing skills can be transferred to ensuring that any policies developed in my professional work are correct.
STEM Ambassador Florence Nightingale	0.5	I volunteered at the annual Florence Nightingale Day. This event showcases successful women in mathematics at various stages of their careers, and displays information about the broad range of possibilities offered by a degree in mathematics or statistics. This year I tutored during the mathematics quiz.	Discussing the mathematical questions with the 16-17-year-olds made me think about how best to communicate mathematical principles in a simple way.	Practising communicating statistical principles with non-specialist people is excellent practice for liaising with clients.
STOR-i Conference 2017	5.0	The STOR-i Doctoral Training Centre conference took place over two fays and included talks from Professor Finn Lindgren (University of Edinburgh), Professor Paul Harper(Cardiff University), Professor Qiwei (LSE) and Dr Steve King (Rolls Royce) as well as STOR-i students and Alumni.	The STOR-i conference highlights up-to-date statistical methodology. Increasing my knowledge of new techniques in extreme value theory, changepoints and statistical learning.	Ensuring that I am up-to-date with the latest statistical methodology allows me to provide relevant analysis for my clients.
Leadership talk with Kevin Glazebrook	1.0	Kevin Glazebrook is Distinguished Professor of Operational Research at Lancaster University. He gave a talk on leadership in academia.	Kevin gave some excellent advice in terms of how to succeed in academia. He gave examples of networking, discussed how to make the most of conferences and gave advice on publishing in both Statistics and Operational Research.	Developing my leadership skills will allow me to deliver a professional service to my clients and develop our business further.
Lancaster Data Science Meetup	1.0	Tassos talked about OpenStreetCab, a mobile service for taxi price comparison. He discussed the open datasets that made the app possible and gave insights on urban transport economics and routing behaviour.	It was fascinating to hear how Tassos has taken an academic project right through to a real-life app. This made me think about how important it is to have a focus when developing a new product. There was also an excellent networking event where I met local data scientists.	My work involves taking academic ideas and bring them right through to the Technology Readiness Level 10. Hearing Tassos's story helped me re-think the importance of including customers at all levels throughout the process.
Personal mathematics tutoring	3.0	For 6 weeks, I privately tutored two students, one GCSE student and one A-Level student.	Revisiting the material for GCSE and A-Level was a great way for me to re-connect with basic mathematical principles. Having to explain these concepts to teenagers helped me rethink the way that I communicate, using many examples.cAlso, having to structure the lessons and set homework encouraged me to reflect on planning.	Being able to succinctly explain mathematical concepts to lay-people will greatly improve my ability to engage with clients.
Arctic Shores Training DataHub	2.0	Qasim and Richard delivered training to me on how to use the DataHub. This included submitting a pull request and the python style guides used.	This detailed training session taught me how to use the Arctic Shores Data Hub to submit queries and pull data. I also learned how to submit a pull request to the Arctic Shores Bitbucket account.	My ability to use the DataHub well directed affected to the quality of the data analysis that I can provide Arctic Shores. Being able to quickly navigate around the Data Hub and find and extract meaningful data is a huge part of my job at Arctic Shores.
Arctic Shores Training Measuring Personality	2.0	Emily Boardman delivered a training course on psychometrics and the way to measure personality traits.	Having a good knowledge of the personality traits ensures that I can consider these whilst developing models for Arctic Shores. It is also personally interesting to understand my own personality traits better.	As a data scientist at Arctic Shores, I will have to develop statistical models which predict personality. Having an in-depth understanding of the traits that we are trying to model is essential to the development of the model, and therefore the service which Arctic Shores provide.
Lancaster Data Science Meetup June	1.0	Matt Barnes talked about data science methods used to analyse air quality data.	Matt gave an excellent talk about analysing air quality data. He highlighted a number of specific challenges with collecting air quality data which had not occurred to me before. We also had an in depth chat about using API's. Since this talk I've been enthused to investigate Open APIs and see whether these can be utilised in the systems that we are developing.	I am currently investigating methods to integrate open source weather (particularly API systems) with our client data where required to create deeper, more improved systems.
Reading Reproducible Research	2.0	I worked through the excellent tutorial by Karl Broman on Reproducible Research.	I aim to use a workflow which meets the gold standard for reproducibility. This guide gave me some great tips and ideas.	By providing a reproducible workflow, it increases the reliability of the results provided. Also, if the data changes it allows a quick re-analysis.
RSS Video Data Science and Statistics different worlds?	1.0	A panel of data scientist and statisticians discussed the role that data science plays within the field of statistics. They covered topics such as if the relevance of statistics been diminished because of new technologies which need a radical new approach, and if data science about ‘getting the job done’, and statistics is about the deeper scientific understanding.	I really enjoyed this discussion. Often, people ask me if I identify as a statistician or a data scientist, and what the difference is between the two. These talks helped to clarify my thoughts and provided interesting angles on the discussion. It got me thinking about the similarities and differences between the disciplines. It also encouraged me to reflect on the skills I have as someone who sits in the middle of statistics and data science.	The personal reflection helped me identify the strengths that I have due to my abilities in the intersection statistics and data science, and skills which may need development. Targetting these skills will ensure I provide a good service to users of my service.
Manchester R August	1.0	The presentations were on Hypothesis testing (Catherine Leigh), Analysing running data (Kaylea Haynes) and {shinytest} (Chris Campbell).	Although I use shiny regularly, and am aware of the importance of testing code, I had not heard of {shinytest}. Chris introduced the R package {shinytest} and demonstrated how easily tests can be implemented into shiny apps.	Learning about {shinytest} has inspired me to include testing inall shiny apps that I develop. This should help flag up any bugs quickly, creating a more dependable service for our users.
RSS Stats Ambassador Panorama	2.0	Assisted Stephen Walsh, Assistant Producer for Blakeway creating a Panorama programme Plane Drunk	Liaising with Stephen, understanding his question and explaining the consequences can to him helped develop my communication skills. Working with a TV company gave me the experience of working to a tight deadline.	Developing my communication and time keeping skills are keep points of providing a good service in statistical consultancy.
Reading R4DS	4.0	I worked through the first half of the R4DS book in Autumn 2017. This book focuses on teaching how to do data science with R. I learned how to use key tidyverse packages such as {dplyr}, {tidyr} and {ggplot2}.	This book is the gold standard for data science in R. Working through the book taught me new skills, particularly in visualising data using {ggplot2}.	Having strong R skills are essential for both analysing client data, but also visualising their data in a clear, intuitive way.
Reading Elements of Statistical Learning	4.0	I worked through through the following chapters, Linear Methods for Regression (3), Unsupervised Learning (14) and Random Forests (15).	As a statistician, it is important to understand the mathematics behind techniques such as linear regression, unsupervised learning and random forests. I focussed on these chapters are they are the most important to my current work with Arctic Shores, however, I would like to work through the entire book at some point.	Having a good understanding of these modelling techniques allows me to develop models for clients whilst being aware of possible pitfalls and biases.
RSS Lancashire Talk David Spiegelhalter	1.0	David Spiegelhalter spoke on Trust in Numbers. By considering the ‘pipelines’ through which scientific and political evidence is propagated, David discussed the possible ways of improving both the trustworthiness of the statistical evidence being communicated, and the ability of audiences to assess the quality and reliability of what they are being told.	David's talk presented a number of excellent examples which I can use in my role as an RSS Statistical ambassador. The pipelines that he discussed has encouraged me to take responsibility for the way that my research is communicated. I will do this by ensuring that all press releases are accompanied by more public-friendly versions, there by helping University press offices do their work.	In my role as statistical ambassador for the RSS, David has provided me with many excellent examples which I can share when educating members of the public.
Psychology Summit	6.0	Work summit in London, introducing me to the company, how our department works and reviewing our plans for writing papers in 2018.	This session was invaluable to introduce me to the science department of Arctic Shores. I learnt about how the department operates, what their outputs are and who does what. We discussed which journals and conferences are key in psychology, and the role of the British Psychological Society.	Having a good understanding of how my department works is essential to quickly providing good quality analysis for our clients. Targetting the correct conferences with papers provides our clients with confidence that our tools have a strong scientific basis.
The Tidyverse Style Guide	1.0	I worked through the style book and made notes to develop my own personal style guide for writing R code. Good coding style is like correct punctuation you can manage without it, butitsuremakesthingseasiertoread. This site describes the style used throughout the tidyverse.	Forcing myself to abide by my own style guide ensures that I am constantly good standard code. It also makes it easier for other data scientists to read my code and provide feedback. I intend to create a Digital Rail code guide.	Having a clear style guide makes better code and, therefore, better analysis. It also ensures that any code handed to the client is easy to understand and use.
Stylish code	1.0	I learnt about using the lintR and styler packages to stylise my code.	It's important as a data scientist to be up to date with the new R packages. These two, in particular, have become part of my daily coding practice, making my code cleaner and more concise.	Having clean code is essential to good coding practice.
Manchester R November	2.0	An R meetup in Manchester. Presentations were on Neural Nets (Martin Eastwood), Reproducibility (Tania Allard) and R in production (David Springate).	Martin gave an in-depth introduction to neural networks. Although I was aware of this technique before, something about how he explained everything really made something click for me. I feel my understanding of them, and confidence to use them has increased. I've thought about reproducibility a lot and therefore most of Tania's talk did not convey new information. However, she did introduce me to workflowR, and R package which I am keen to try out. Using this will help my improve my whole workflow with in R. I have also decided to investigate using docker. David's talk really got me thinking about using R in production. Although I think for the work that I do, it is still the right solution, He highlighted some work around that can be used to help R be more stable in production. These technologies included Packrat, install_github(“hadley/strict”) and Data.table	By implementing the knowledge I have gained through this session, I will aim to provide a client service which meets gold standards for reproducibility.
Mango Data Scientist Radar	1.0	The Data Science Radar as a conceptual framework that allows users to explore these character traits in more detail and to uncover their own data science profile type. I scored as follows, Programmer (6.2), Modeller (5.4), Communicator (4.3), Data Wrangler (4.1), Visualiser (3.8), Technologist (3.4)	I have a tendancy to try and be good at everything rather than specialising in specific skills. Looking at my radar has reminded me of the skills which are less important in the roles that I do. I think I should focus on maintaining my Modelling, Programming and Communication skills rather than trying to keep on top of all the new different technologies (Docker, Hadoop etc.)	By identifying the areas in which I should focus my development, I can tailor my long-term data science development to best support my existing skills and my clients needs.
Women in Data Conference	5.0	A one day conference on Women in Data. Major topics included AI, GDPR and self-promotion	This event introduced me to many role models in my field, and really boosted my confidence. I learnt some great techniques on how best to communicate and present myself in a business setting. One major theme of WiD2017 was taking control Athina Kaniouva said, “if you want a promotion, you have to ask for it, don’t wait for it.” Håkan Nyberg, CEO of Ikano Bank, said that women have a “golden opportunity” to help, promote and sponsor each other. There was also a in-depth discussion into GDPR which increased my understanding of the upcoming legislature.	Thanks to my better understanding on good GDPR practice, we can provide a service which meets users expectations around data protection.
Arctic Shores Training Security	1.0	Security training delivered by Jorn on best practice. Included choices for passwords, two-factor authentication, public wifi and laptop encryption.	Jorn's presentation taught me how to ensure that Arctic Shores data is kept secure. As part of the session, I set up two-factor authentication on my mobile and encrypted my work laptop. We also discussed accessing the DataHub over public wifi.	Arctic Shores stores some personal data of the users. Therefore, we have a duty to follow UK guidance and best practice to keep their data secure.
Arctic Shores Annual Summit	2.0	Presentations from different employees of Arctic Shores. There were talks on Upgrades to the DataHub, the History of Psychometrics, Inspector Help and Automating interview questions	The variety of talks from different departments of Arctic Shores made the session very interesting and beneficial for me. Working as a lone statistician in a psychology company, it's vital to connect with the other departments. The talk from the DataHub about Inspector Help highlighted functionality with the DataHub which I did not previously know about. The Psychologist provided a history of psychometrics which was fascinating and helped me understand where the development of our tests have come from. The client services team showcased a new product in development which I will be working on in the near future. I also gave a talk on my work exploring the relationship between traits and levels which was apparently very useful.	Understanding how the company works, and what projects other departments are working on is vital to providing a great service to our clients.
Training Writing Aptitude tests	8.0	Training material delivered by Assessment Day. Covered how to develop numerical aptitude tests, including the writing of questions and assessing the difficulty.	Over the next couple of months, it is my job to develop an aptitude assessment from scratch. As I do not have any formal psychometrics training, it was essential to seek training from experts. I've developed a good understanding of how to develop these tests, from writing good questions to analysing the difficulty.	Being trained in aptitude tests, allows me to write a good quality assessment. Having an accurate assessment will allow our clients to ensure that they employ the best possible candidates with confidence.