Dear all,
I am very much enjoying following the Applied Epi Intro R course with 10x 3.5hour sessions. I think the instructors are great, and it is really nice that there is a lot of practical work and interactions in the course. We are starting week 4 of the 5 week course tomorrow.
My question to the Applied Epi Community is this: Does anyone have good advice for how to continue learning and improving your R skills after we finish the intro course?
I imagine that it will be key to continue to practice and working on applying the new R skills in my daily work. In my case it means practicing using R to create regular outbreak reports, using R to create regularly produced tables etc.
I am wondering if some of you would like to share your thoughts on how you plan to continue using R after finishing the intro course?
I am considering booking time in my work calendar - perhaps 3 hours a week, to practice my R skills and apply them to concrete things in my work.
Anyone else have ideas for what you plan to do after the course to continue becoming a better R user?
Kind regards,
Stine Nielsen (epidemiologist at the department of infectious disease epidemiology and prevention at SSI in Denmark)
I’m glad to hear you are having such a great time with the course!
I think your plan to apply R to your professional work is a very powerful way to continue your R learning. I did this exact thing when I first began working as an epidemiologist and it forced me to learn new things. Joins using dplyr and visualizations using ggplot are examples of things I learned early on as they were necessary for my work.
Another thing that I’ve found useful is reading books on R, many exist online and it may be beneficial to create a reading list here in the near future.
Finally, I have found following other R users on Twitter to be very beneficial. Jenny Bryan and RStudio Tips are examples of accounts I follow.
Hi Stine,
since we are doing the same course I can’t tell you yet if my strategy will work long term. I work with excel on an almost daily basis and after each module I try to apply the learned things to my own dataset. This is for me very important otherwise I forget the code really quickly. Applying it to my own data is a very slow progress (like 3 hours to convert one date variable to the variable I wanted), but it was a happy moment yesterday when I finished my first taylor made epicurve.
My plan is to replace the work in excel with R as much as possible regularly. Even though that takes more time in the beginning I am sure it will reduce my workload long term by automating analysis and reports.
Kind regards,
Raskit
When I first started learning R, it was with the intention of performing a full analysis on my own dataset through R - I found having that focus on one specific analysis project really helpful, as there is so much out there that it can sometimes be a little overwhelming. I would maybe select one or two projects in the first instance, which you haven’t started yet and then develop and execute the whole analysis plan in R. You can use the materials from this course as well as the Epidemiologist R handbook for guidance, and post specific questions here if you get stuck or want a deeper understanding of something.
There are a couple of considerations when doing that which it is useful to plan for:
The first time you do this it is going to take longer than normal, so pick a project that has a long deadline and where some extra automation with R would really add value (over your previous approach).
If you are working with a team / other people will need to share / run / take over management of any scripts you produce, it is worth getting the buy-in of the whole team to do the project in R in advance and make sure there are enough people familiar with the code who could adapt it if necessary, when some future need changes (think transferability, sustainability and succession - who inherits the code or runs it when you aren’t there).
Do code reviews with the other people in the team and also include stakeholders who don’t know R if necessary (they have a different role, to check that general requirements are met and also perhaps to beta-test it and provide feedback on the output). For people who do know R, it is a chance for them to become familiar with your code, where are the critical points that might need adaptation in the future and to learn from each other on the coding approach taken.
If you are working in a department where there are no or very few other R users, it can be very useful to get buy-in from the lead of that department to start a process of onboarding people with R. That could include people from the department doing a course like this together, then arranging weekly code reviews to discuss their own code afterwards, as well as allocating time and people to transition all the regular outputs to R. Applied Epi’s very first R intro course cohort from Public Health Wales implemented this system and might have some advice on how to go about this.
This is actually a very great question. I can share my experience few years ago, when I started learning R and how I progressed so fast than I thought.
So the biggest secret of learning R is to continue to use R in most of your daily data management and analysis activities. you must shift immediately in to R and don’t be embarrassed by the first few reports you might generate while you try to improve them from day to day. The first tables and plots I shared for my first SitRep using . Rmd was quite primary and I was not happy about it. I continued to work on it and the next day I improved few things. on the third day again some improvement and I continued. Know what, finally I was addicted to R and could not stop coding.
coming in to technical assistance, there are several R communities that you can access and ask you questions, you will get support quickly usually. This applied epi community is one of them. I found the reprex function from {reprex} package very helpful, when I wanted to post to stack overflow and people reply to me in a day or two. boom!
I also like to download and study peoples code form their github address. If you google some codes and you find link to github, give it a try and it helps. There was also a time I mailed authors to explain some of their codes. Desperate time!
One big secrete I like to do, I google the package documentation and study what are the functions in that package. Simple try go to google and search for “rio package Rdocumentation” rio package - RDocumentation. You will be amazed by the swiss army knife , but open it and scrolling down you will see the functions like p_load and what they do. Give it a try.
**Biggest advice: The more you interact with R the more you become perfect, and the less you interact with R the more you forget how to code and get frustrated **
, but open it and scrolling down you will see the functions like p_load and what they do. Give it a try.