⌛ Our Selection Process

HOW WE SELECTED THE CITIES IN OUR OPEN DATA REVIEW

Dataset Selection Criteria

The starting point for this project was to determine whether the was sufficient open emergency call data available for us to prototype a system for standardizing 911 calls.

A 1-Day Experiment

Our first attempt to answer this question happened during the 2021 National Day of Civic Hacking. We simply provided 105 volunteers with a list of the Master 911 PSAP Registry and encouraged them to pick a city that had personal meaning to them. During that 1 day event we reviewed 384 individual cities and established that there was enough 911 data available for us to begin experimenting with ways to aggregate and analyze multi-jurisdictional data.

Refining Our Criteria

Our next step was to perform a landscape analysis to determine which datasets were suitable for our standardization experiments. We reviewed cities that met one or more of the following criteria:

• Top 100 most populated cities. Our experiment on National Day taught us that large cities are more likely to share data publicly. The cities are also likely to be easy to cross reference with detailed census or demographic data.

• Brigade cities. Since one of Code for America's strengths is its ability to mobilize volunteers and collaborate with our local chapters (aka "brigades"), we wanted to know which brigades existed in cities that also shared open call data.

• State capitols. In order to ensure that at least one city from every state was reviewed, we also added all the state capitols to our search list.

Our Final Sampling

After reviewing our selected cities we identified 134 cities that provided open data to the public.

These 134 cities made the basis of our "Open Data Review", a landscape assessment wherein we reviewed these cities across 27 different dimensions. During the Open Data Review another seven cities were eliminated.

This left us with 127 fully-reviewed cities. The landscape analysis data for these 127 cities are available as an open dataset on Data.World and Zenodo.