Adam Baker

Selected product design & product management work.

openFDA 2014

A new standard for open data and public health

open.fda.gov

The query explorer teaches people how to use the API, and shows them insights in the data. Here, you can see who tends to report adverse events associated with acetaminophen.

openFDA 2014

A new standard for open data and public health

open.fda.gov

What it is

Created under the vision of Dr. Taha Kass-Hout, openFDA is an open-source, search-based suite of fast, contemporary APIs for important drug, medical device, and food datasets. The entire project is online on GitHub, the data are open (CC0) worldwide, and it provides unprecedented open access to health data about drug and medical device adverse reactions, pharmaceutical product labeling, and product recall information.

What’s special about it

The design reasonably serves a challenging dual purpose: Interactively teach people how to use the API, and communicate and allow non-experts to learn from the data themselves. Indeed, this is in many ways the only place that this important public health data have been published in a way that highlights insights. Thus while openFDA’s audience is developers, the design extends an olive branch to the general public. This project set a new standard for government open data, thanks to the hard work of Dr. Kass-Hout, Sean Herron (a Presidential Innovation Fellow), and the elegant work of our FDA and Iodine team members.

Role

As part of the Iodine team who built openFDA under contract to FDA, I co-designed and documented the API endpoints themselves (again, the first time the data have been so thoroughly documented), and designed and built the website, including the interactive query explorer.

HTML5 + LESS + JavaScript

Looking back

Thankfully the project is still evolving, so there’s as much looking forward as back, but I’d say that we only went half-way with the interactive query tool. A better version would do more guiding, especially around query construction, and also showing people the bounds of the data (for example, somehow showing a list of all the possible drug adverse reactions). These are complex data that take a while to wrangle and understand, and we need to continually improve the documentation and tools to help people get the most out of them.