How to publish annual datasets?

Question

I work in the open data team for a large city. We have a number of datasets that are produced annually, such as budgets. I'm trying to choose between different options for publishing this data:

Single, expanding dataset

A single dataset called "budget" with a year column. I fear that this increases the effort for the consumer, who probably has to filter to get the year they're interested in. Also, sometimes it makes the dataset more complex, as the 2015 budget contains 2014 actual spending, 2015 budgeted spending, 2016 projected spending etc.

One dataset per year

Alternative, we could have "Budget 2015", "Budget 2016" etc. This creates some logistical challenges, such as the difficulty of automatically creating (as opposed to updating) datasets. Also, it's not clear how you would answer a question like "Is this dataset updated regularly?"

Both?

Perhaps the right option is a single expanding dataset, with filtered views on it?

Is there any guidance here? Best practice? The platform we're using is Socrata but I'd prefer platform-agnostic advice.

Answer 1

I think having separate datasets for each year would probably be the easiest for users of your site. I work with lots of government data portals, and the norm tends to be to have a new dataset each year. It does add a bit of complexity technically, in that a new dataset has to be created each year, but it is much easier for the average citizen to browse this way.

That said, if you make it very easy to find the filtered views, having one master dataset with annual filtered views would be a lot more friendly to any developers working with your API. But that approach requires that you be confident that the format of your dataset will not change from year to year.

I guess the approach you take depends on which audience is more important to you, the general public or developers. I think either approach is valid and won't cause significant pain for one audience or the other.

---

Since you are on Socrata, have you looked at their Open Budget platform? It provides an extremely pleasant end-user experience for browsing budget information. You can check out their demo portal for the fictional City of Evergreen, or, for a real usage example, check out the portal for Montgomery County, MD. Note that I'm an end user, so I'm not familiar with pricing or ease of setup for Open Budget.

Answer 2

If you're interested in seeing how some other large cities are publishing their data, here's a few examples of large cities' organization pages on data.world. Hope this helps!

https://data.world/san-diego https://data.world/cityofaustin