We are a group of research students studying Applied Science and Informatics at NYU's Center for Urban Science and Progress. We were tasked with creating a tool to help make it easier to use 311 data in new contexts. What we found was that it's not working with the data that's difficult, but the interpretation that makes using 311 data a reliable addition to research. For example, categories of complaints and geographies change over time. What does this mean for the reliability of the data? If something is recorded differently depending on the way it's reported, is it still valuable to include as texture in research?
While we can't fully unpack the inconsistencies in this user-generated data, we wanted to highlight questions to keep in mind about the data, while providing examples of how it can be useful.
Our report is available on Authorea, and you can view our code on our GitHub page.
Have something to add? Contact us
We can't wait to bring more cool features, help us get there! Support us through Patreon