write script to create database of KCDE predictions from our public data

[nickreich]

Predictions live at https://github.com/reichlab/ssr-influenza-competition

[nickreich]

Directly readable from https://raw.githubusercontent.com/reichlab/ssr-influenza-competition/master/inst/submissions/EW01-KoT-2016-01-18.csv

[qbazinet]

Just uploaded a file called scanned_weeks.R that should do exactly this! It divides each week into two dataframes (for the two different sets of data within each document), and appropriately names all columns. It's a bit inelegant, but it should get the job done for the moment.

[nickreich]

It works for me! Nice first pass getting some code that works and creates simple readable data. Now the trick is to functionalize the code so that it could be applied more generally. Ideally, we'd have a script that could run and create 1 or 2 data.frames that could then be saved. I'm thinking something like this process:

• [ ] write a function that takes a url and returns the two versions of the dataframe for a single week
• [ ] write a second function that takes a list of urls (one for each week) and returns two large data.frames
• [ ] in a separate R file, source the script with the two above functions, then create the two data.frames and save them in a data directory