Dev sprint 1&2: Setup

Learning goals

• When do users want access to the API vs data downloads?
• What is an acceptable size for a data download?
• What are important details to include with data downloads?

Participants

• 18F team conducted 4, 1hr remote interviews with data experts working in the private and public sector

What we learned

Expert analysts will use the API if they want to build something and downloads if they want a static data file. Consumers want to quickly scan the contents of a download to understand what’s included and then refer to documentation, such as a glossary of terms and a schema, to answer specific questions. Users have confidence in the data when they know the source, methodology, and have access to a point of contact.

Experts expect files ranging from 1.5 – 2 gigs in size, but older versions of Excel truncate files with more than 65,536 rows. Continue to explore the best way to shape csv files. The CDE needs to clearly explain what the different datasets are (SRS, NIBRS) and how they will be made available (in what formats). Dynamic feedback about dataset usage, updates, and common questions would be valuable to users.

• Downloads and previews help users understand what is in the data and are useful even if they ultimately will use the API. API=Build something, Download=Data Users want to understand the source of the data - who is providing it and why?
• Points of contact and ability to provide feedback are important
• Need to more clearly explain what the different data sets are (SRS, NIBRS) and how they will be made available (in what formats)
• Users want to be able to preview, or quickly scan the contents of a download to get a “sense of the data”, and then refer to documentation, such as a glossary of terms and a schema, to answer specific questions about how the data was collected and how it may be used.**
• While experts prefer more, rather than less data, it’s easy to get lost in large datasets if you don’t know what you’re looking for.
• Users expect to do some manipulation of the data in order to fit their needs.
• Feedback about dataset usage, updates, and common questions would be useful to users
• Things that failed- type of data design and wording needs refinement, “timeframe” option for slicing,

Sprint 2 recap

Next steps

• Build core components of “explorer view” - basic visualizations & navigation of data
• Validate & refine design approach
• Begin developing content to support the data & schedule a workshop with CJIS to develop a longterm approach for content management
• Make decisions about file size to support as data downloads

Prototype tested

https://gsa.invisionapp.com/share/TC9VBXR8K#/screens/212669091