Chemistry is about trial and error. Experiment is necessary for discovery. That said, the better we analyze past experiments, the better we can predict what to try next. That's where the Dark Reaction Project comes in.
Haverford has large collections of old experiments performed by students over the years. I aimed to use those data to help chemists decide what to pursue next.
The analysis and visualizations from this project are in use on the Dark Reaction Project website, and were featured at a research symposium and discussed by a chemistry professor at a conference of the American Chemical Society.
Many schools and universities collect data on past experiments and whether they succeeded or failed. To some degree, future success and failure will resemble those existing data.