[okfn-labs] Policy for multiple data files per dataset for data.okfn.org

[Rufus Pollock]

Hi,

I've just added a US House Prices to http://data.okfn.org/data:

<http://data.okfn.org/data/house-prices-us> <-- catalog entry
<https://github.com/datasets/house-prices-us> <-- source data package

It raised an important question: Should this have just one data file
or many? (Just one data package or many)?

To provide some context:

- At the moment, I've only added one data file (national index,
annual) but I could have several and current setup means data is
"missing"

- The source data comes in several files (specifically: national index
(quarterly) and 20 city index monthly). I feel I should include as
much data as possible and as granular as possible but at the same time
want simplicity and consistency.

- My approach so far for other time series has to been to go with the
annual series as the default hence the choice here but in some cases I
have included the more granular data as separte files

What's the best approach here? Should we allow multiple data files or,
instead, have multiple data packages (named e.g. house-price-us,
house-prices-us-city, house-prices-us-qq (quarterly))?

Regards,

Rufus