[wdmmg-discuss] CSV schema of user uploaded spending data

[Anna Powell-Smith]

Hi Stefan,

> based on the Technical Format (http://wiki.openspending.org/Technical_Format) I have written a user guide to creating a conforming CSV file for user uploaded datasets.

Really good to see this!

> This is the first iteration and a simple loader that can read it will be implemented. Please have a look and give feedback.

The technical spec looks very clear. Two questions:

1. Should 'currency' be required? - if not, specify the default currency.
2. Clarify whether name and description can be the same, for datasets
with little descriptive information.

Just a few things it may be useful to clarify for users, having worked
with various spending datasets in the past:

1. Datasets should not include aggregations - we are only interested
in actual transfers of money. (People sometimes try to be helpful by
including aggregations.)
2. Datasets often include both budget and actual spending - if so,
must specify what kind it is.
3. If the dataset includes hierarchical categorisation information,
clarify whether the users should split out each level of
categorisation into a separate column.

There's obviously a limit to how much you can specify in advance, and
I think you have a good balance here, but those were the most common
issues I came across.

cheers
Anna