I certainly agree with the point, but I would go further still. I think of opening data as a kind of ladder we must climb. Data can be:
- Requestable – the public has a right to the data, but must specifically ask for it, e.g., through a FOIA request;
- Findable – the data is publicly accessible, for example, a PDF posted on a county website;
- Usable – the data is machine-readable, e.g,, available as JSON or CSV;
- Interpretable – the data includes meta-information that allows a specialist or committed investigator to understand the context and limits of the data; this is what we really need to ask of data providers;
- Understandable – the data is presented in a form that allows an ordinary citizen to understand it, e.g., effective visualizations or narratives built around the data;
- Useful – the data is presented at a time and in a form that is relevant to the target audience and makes it easy to actually use, e.g., restaurant inspection grades presented as part of a Yelp record.
As the GovLab post points out, making information machine-readable doesn’t go far enough. In this framework, it’s only halfway up the ladder. Data only becomes accessible to ordinary citizens at the 5th rung and, by definition, only really matters for the majority of people when you take it all the way to the top.
Of course, in general this is really hard. What we’ve done with the Community Budget Explorer so far, for example, only makes it to the 5th rung, and we could even do better there. But hard or not, this is the standard we need to aim for.