Generating code for weekly topics

[choldgraf]

I think it'd be useful if, for a given topic, there was a set of code that would:

1. Do the analysis manually.
2. Do the analysis with a readily-available package.
3. (optionally) show number 2 for both Python and R

Would also be good if there were a publicly-available dataset that could be revisited on a weekly basis.

Thoughts?

[rlbarter]

I think it would be great to have the blog post walk through an example based on publicly available data - this has been suggested by others too.

Are you suggesting that we have the same publicly available dataset each week? Or just that we should compile a list of the publicly available datasets that we use each week?

[choldgraf]

I think it would be great to have the blog post walk through an example based on publicly available data - this has been suggested by others too.

Cool - I could imagine this being converted into a little e-book or something if there's enough content generated. Always good to think about how you could publish or otherwise "officialize" the things we're doing :)

Are you suggesting that we have the same publicly available dataset each week? Or just that we should compile a list of the publicly available datasets that we use each week?

There's an argument either direction. Groups like data carpentry do their analyses on the same dataset so that you don't have to spend time re-orienting students each lesson. They usually find a large-ish messy dataset to use. On the other hand, it'd be cool to see lots of different kinds of data. So maybe just see what the team thinks about it.

[rlbarter]

I like the idea of having many different publicly available datasets - some data are more relevant for some methods than other. It can be time consuming to find relevant datasets though. We should discuss on Thursday :).

[choldgraf]

Cool - yeah either one is good but I think it's important to make sure the data is interesting and available for folks at least :)