How should we choose our datasets for our City data census?

Hi all,

We’re about to fill out the South African City open data census and we’re wondering how we should choose our datasets.

  • the default choices seem like good ones – is there motivation for those? Are they a good set to start with?
  • if we use the defaults, we could compare ourselves against other cities internationally. Is that a reasonable consideration or is it wishful thinking?
  • what factors have others considered when choosing their datasets?

Here’s our current set:

  • Real-Time Transit
  • Annual Budget
  • Expenditure (detailed)
  • Election results
  • Air Quality
  • Transport Timetables
  • Public Facilities
  • Crime statistics
  • Tenders awarded
  • Food safety inspections
  • Traffic accidents
  • Building permits
  • Service Requests
  • Business Permits
  • Business Listings


In Australia, at the State level, this is the process I followed.

At the City level I reviewed legislation to determine what Councils must publish but these datasets weren’t the most useful so others came up with a list of more useful data. With around 600 Councils in Australia, consultation is a challenge.

To make it easy for Councils to publish, so data can be combined and compared, Open Council Data are developing standards.

1 Like

Thanks Stephen, this is useful, particularly the G8 charter reference and your approach.

We’re building a list of candidate datasets here:

This looks great! @asmenokfn and I want to get started with crowdsourcing input from Swedish users of what data that would be great to measure. This looks like great inspiration, @asmenokfn will add a suggestion for the Swedish Local Open Data Index and census. We also need to figure out how to crowdsource the info in the best ways since there are around 300 local authorities to review.

A word of caution… some dataset sets that were added to Australia’s Local Open Data Census are managed by the State’s (e.g. Ward maps that indicate local council electoral boundaries). In this situation, and with around 600 councils, we need a way to do bulk entry into the census.

Research the dataset ownership before deciding on the inclusion of a dataset in your Local or Regional Open Data Census.

1 Like

Hi, @longhotsummer, this list is really useful:

Would you consider making another version of this spreadsheet so that other city datasets could be added? It would be great to have them in one place for easy comparison.

Hi @ewan_klein – we aren’t maintaining that spreadsheet at all, but you’re welcome to make a copy of it and add any other countries you’d like to see.

It could also be something that OKFN could automate pretty easily if there’s a compelling use case.

1 Like

@ewan_klein @longhotsummer please involve @asmenokfn in creation of this. Would be great to automate. Also maybe separate topic - how can we automate as much of the census work as possible? Please share your ideas. What we’ve been thinking is to use We’re compiling one for Swedish authorities to automate e-mail for as FOI request. Create new topic if you see it fit. :slight_smile: