Astro: A Star Map for Navigating CMS
Need to organize your space? Call Marie Kondo. Need to organize your data? Call Ben Damman.
A self-taught software engineer, Damman has been collaborating with the OIT Front Office to design a map of all the knowledge contained across CMS. The project, codenamed Astro, will expedite the work of CMS staff and prepare the agency for the implementation of artificial intelligence (AI).
“The employees at CMS are knowledge workers,” explains Damman. “The most valuable thing we do every day is to think.”
When information about healthcare, policy, or operations is organized and easily accessible, CMS employees are able to devote more energy to using their brains.
“If the information you need is at your fingertips,” says Damman, “you can make informed decisions faster.”
Damman stresses that Astro is not a repository for information, but more of an index. It will help you navigate across the wide expanse of knowledge about CMS data and documents (within our OIT systems), knowledge about the CMS enterprise (held within our people), and knowledge of interest to CMS (held elsewhere).
If you are looking for a person with specific expertise, or perhaps a specific document, it will point you in the right direction.
The interface will be web-based, so, like on the World Wide Web, related elements of information will link to one another.
Damman compares it to the Federal Register, which offers the history of all U.S. legislation. The registry itself doesn’t house every relevant document, but it does link to them.
It also informs users about the different versions of a bill, including “metadata” that details how, when, and by whom documents were updated. Astro will similarly inform users about how CMS data has changed over time.
Well-defined data is not only ideal for those who must access it, but also for interaction with machines and algorithms.
“For AI, data is way more important than computation,” says Damman. “Like with all software, garbage in, garbage out. AI loves highly structured data. It gives it more signal. The more information you can feed into a machine learning (ML) model, the better the resulting solution will be.”
Damman’s experience building civic tech products includes a tour of duty in the U.S. Digital Service, where he worked on login.gov and wrote GI Bill software for VA.gov. Prior to joining the White House, Damman was a software engineer at Apple. He also worked on California’s “motor voter” initiative where residents are automatically registered to vote – unless they opt-out – while they are obtaining a license or ID from the Department of Motor Vehicles.
This background in building complex systems has imprinted upon Damman the importance of sharing knowledge.
“So much of this is about communication,” Damman says. “The more specific your information, the more actionable it becomes.”
A good map improves coordination, facilitates cooperation, removes blocks, and encourages exploration.
Astro, which is in the very early stages of prototyping, is itself an act of exploration. The current development plan calls for the code to be open source and available to all federal agencies.