Standardized filenames

[jsoma]

It could be helpful to run with a consistent naming convention like country-datasource-YYYY-MM-DD.csv, e.g. liberia-case_reports-2014-09-29.csv - I think it might make sorting and browsing a little easier.

It might also help with #37 so that guinea-report-2014-09-29.pdf could sit next to guinea-report-2014-09-29.csv and you'd have a better idea what still needs to be digitized.

[cmrivers]

Totally agree, wish I had standardized both the file names and the variable names. At this point though I'm worried that changing it would be burdensome for users, e.g. might break scripts.

Anyone else want to weigh in?

[chendaniely]

burdensome for users, e.g. might break scripts.

do you know how many people are actually using those data? and potentially how many people will be affected?

In the long run, it may actually be more beneficial to standardize names.
Since sorting and tracking will be much easier to do when things are properly sorted.

I want to say nicely named files are probably better for scripts anyway, compare this to trying to parse the current file name.

If you've ever worked with census data, the fixed file naming convention (including how many characters) is a godsend.

[chendaniely]

you can always supply a renaming script :p

[cmrivers]

I don't know exactly how many people use them, but I don't think it's a trivial number.

I'm leaning towards standardized renaming, but I think we should leave open this issue until maybe Monday the 13th to give people time to comment. At the very least I think we can sed out the spaces in the filenames, since that annoys even me (and I'm the one who put them there...).

[chendaniely]

may I suggest 2 digit numbers (e.g., sept 01 vs sept 1)

[donpdonp]

Standardizing the filenames would help anyone who will be importing all the data into another tool.

/countries
  /liberia
    2014-09-29.csv
    2014-09-19.csv
  /sierra_leone
    2014-09-29.csv

If there are really different categories from 'casedata', a third level of directories might work, too. The date should follow iso8601.

A filename reorganization is a step along the way to get all the csv into a single data source such as a sqlite database.

[samccone]

👍 @donpdonp