Home  -   Research  -   Artwork  -   About Me

Google Product Visualizations

Visualizations that augment high-traffic products and services present unique challenges and opportunities.

My team has launched work on:

  • Search
  • Translate
  • Sheets
  • Analytics
  • YouTube
  • Google+

At Google I've had the opportunity to work with products used by millions (sometimes billions) of people. Designing visualizations that fit into the rapid-fire, casual usage of everyday applications requires an intense focus on simplicity.

The image above shows the "Explore" panel of Google Sheets, which creates automatic views of spreadsheet data. My group led the research for this feature, and then worked closely with the Sheets team to launch it. In addition to creating new techniques for recommending charts and visualizations, we invented a class of "data verbalizations" which make charts accessible to the blind—and provide summary captions for everyone. The feature has been a success, and has been followed by much related work both within Google Sheets and (perhaps inevitably) Microsoft Excel.

At left is a table of back-translations which we created for Google Translate. This feature solves a perennial problem: if you're translating a word into a language you don't know, how can you be certain you've gotten the right sense? Users often would perform many manual back-translations; this table automates that work, and provides language learners with helpful material. A key challenge in designing this display was to keep it as simple and unobtrusive as possible, given that it appears for every one-word English translation query.

For Google Search, my group created a new type of "dictionary"—a special box that appears when users make a query such as "define mellifluous." Users can (and in practice often do) expand a basic definition to see additional details, presented in graphical format. The idea is to take advantage of the fact that on the web, we have infinite space, unlike the cramped world of a traditional paper dictionary.

We worked closely with the Search team to design and implement this feature, and performed extensive user testing (video recording, eye tracking, etc.). The charts, for etymology and use over time, are highly pared-down versions of the original mocks and prototypes: a key lesson from this project is how much restraint matters when designing for an extremely high-traffic application.