I have been asked to host an Introduction to R workshop for the analyst team of the company that I work for. It's a large retail company. Currently it's only me and one more person that uses are extensivly. My idea is to follow the R for Data Science book by Hadley Wickham and Garrett Grolemund.
Today the team primarily uses SQL and Excel to do analysis. For statistical modelling one of the analysts has an SPSS-license. The teams amibition is to move more of the analysis to R to excel (without Excel) in Data Science. However, their ambitions are pretty scattered. Some of them want to develop skills in Machine Learning, others in A/B testing and some in forecasting. So I'm trying to figure out a way to keep everybody happy.
For the setup I plan to use RStudio Cloud.
As for the curriculum I have proposed 4 half day blocks to cover the basic. The blocks will be spread out during two weeks.
Block 1: Introduction to R
Here I will focus on dplyr and tidyr to tidy and transform data Block 2: Visualization in R
Mainly ggplot2 but here I'll also introduce Rmarkdown and how to build reports Block 3 Databases in R and case study
I'll introduce some best practices on how to work with databases in R, how they can use dplyr as a front end to the SQL-server. Furthermore, here I'll introduce a case study so they can practice retrieving data from the database and make analysis in rmarkdown using the skills from block 1 and 2. Block 4 Statistical modelling in R
I'm not quite sure what to put into this block. My first thought was to do something about A/B testing in R, so that they can further test their current campaigns. However, more people (10 today) have signed up for the workshop and maybe only 3-4 are intersted in A/B testing. So I'm thinking of switching this to something more broad. Maybe just how to do some basic modelling in R using tidy tools for modelling, that can be extended to different areas such as time series analysis.
This is not my first workshop in R, however, I have a feeling there's a lot at stake here and I really want to get the team up running in R because I think it will benefit us hugely.
What are your thoughts about this curriculum or setup? Would you have done it differently? Can you see any pitfalls?
I work in a large university department where we are migrating from widespread us of SPSS and Excel to the widespread use of R. I've been doing a lot of "Intro to R" workshops to help this migration.
Some general comments:
If your audience is eager to learn R, everything is easier and more or less anything you do will be appreciated. If your audience is sceptical about the need for change, and if they're just looking for reasons to dismiss R, then you have to be very careful because they could be very unforgiving.
If your audience has programming experience of any kind, then commands and scripts etc in R will be easy for them to pick up. If they're only used to point-and-click software, you need to introduce commands very incrementally and on a need to know basis. For example, introduction to R workshops that start with defining fundamental data structures can be a real turn off to people who barely know what a data structure is.
Focus first on things that they need to use on a daily basis. For example, people in my group do not use SQL at all, and so I would not go into that as part of introductory workshop set. On the hand, they do statistical data analysis (e.g. regression modelling) all the time, so getting into this relatively quickly and deeply is a priority because it puts them on familiar ground and they can see how R would be beneficial to their daily work. If your audience doesn't do statistical modelling on a daily basis, then your Block 4 might not be the best option. If they're doing more data wrangling and tidying on a daily basis, then maybe going deeper into dplyr and tidyr etc would be a better option.
I think Wickham and Grolemund's book is great, so it's a really good choice to follow.
Also, I'd definitely go, as you are planning to go, via the tidyverse way, not the base R way for everything.
I really agree that it is important to focus on the things they need on a daily basis.
Yeah, that is true for Block 4, I'll have to think about that. One idea is to introduce some simple linear modelling concepts and at the same time introduce broom for making sense out of statistical obejcts and purrr for functional programming.