Work/Ground Truth
Ground Truth
Data tooling for a climate research lab.
- Client
- Stanford Earth Lab
- Role
- Lead Designer
- Discipline
- Product · Research
- Timeline
- 7 months
- Year
- 2022
Tools
Figma · Observable · D3
[ 01 ] The brief
An internal tool for climate scientists to compare satellite measurements against on-the-ground sensor readings. Built with the team that uses it every day.
Researchers were stitching together QGIS, Excel and a homegrown Python notebook to validate satellite estimates. A single comparison took four hours and an unhealthy amount of caffeine.
[ 02 ] Approach
- 01
Embedded with the lab for three weeks. Watched the four-hour workflow happen in front of me, and asked dumb questions until the bottlenecks were obvious.
- 02
Designed a side-by-side viewer with synchronized zoom, a shared timeline, and a difference layer. The four hours collapsed to 20 minutes.
- 03
Worked directly in Observable so the prototypes ran on real data from day one.
[ 03 ] Selected screens
FIG. 01
Comparison viewer
[ 04 ] Decisions
No charts for chart's sake
Every visualisation answers a specific scientist question. If a table did the job better, we used a table.
Reproducibility in the URL
Every view is a permalink. Researchers paste comparisons into Slack and Github issues; the link is the citation.
[ 05 ] Outcomes
- 01
The validation workflow time dropped from 4 hours to 20 minutes.
- 02
Tool was adopted by two partner labs at MIT and ETH Zurich within a year.
[ 06 ] Adjacent work
Next project →
Atlas Archive