ecortens / cirpa-r-workshop-2018 Goto Github PK

From our conversation, we are not as concerned with finding IR-related datasets - but rather ones that will work well for our purposes. (Mostly around the type of variables they have).

This issue is going to be my place to talk about some datasets that we can use. As per Evan's suggestion, I have been looking at the UC Irvine Machine Learning Repository which has a large number of datasets organized by domain and kind of problem.

One potential dataset is the Bike Sharing Dataset of Bike rentals in Washington DC. It has calendar and weather information. and could be used to look at differences in the number of rentals by the characteristics of the day.

We want to add a few resources to the document. This list is a work in progress.

• Add a link to the Data Carpentry R Social Science lesson: https://datacarpentry.org/r-socialsci/

Update the information to match the 2018 CIRPA conference and to add me as a co-author. The file names will need to be updated as well.

(I have already created some commits that do this -- and when I push them, I will close this issue.)

One of the principles that I would like to get across is the idea of visualizing first (which fits in with the current presentation and the R for Data Science book). (Though I will admit that I don't always follow my own best practice here.)

We could add in a discussion of the visdat package, which could be useful. The downside it it's another package to install.