Computational biology projects built with AI
HER
Predicts menstrual phase and ovulation from wearable signals
Built by Amatullah Rampurawala, Sahana P and Vidit Jain
What it does
Most cycle apps assume ovulation on day 14, though cycles naturally run from 21 to 35 days. HER learns from the body instead: a Random Forest model trained on 16 wearable measures, including resting heart rate, heart rate variability, sleep and temperature, predicts the current phase, the chance of an LH surge and the fertile window.
How it works
- Uses phase-labelled wearable recordings from the PhysioNet mcPHASES dataset
- Trains a Random Forest on 16 physiological features, balanced for cycle variation
- Takes daily values or a CSV and returns phase, surge probability and fertile window
Built with PhysioNet mcPHASES · Random Forest · HRV · Python
A learner prototype built on public data, shared with the team's permission as part of their portfolio. It is not a clinical tool.
More projects
HerForecast
A daily readiness forecast across the menstrual cycle
See the project
Solace
Morning guidance for women in perimenopause and menopause
B. Melissa Judith, Bhavika Pawar and Rena Pawah
See the project
FlashReady
A morning heat-sensitivity forecast for hot flashes
Rena Pawah
See the project
Build your own in the PGP
Six months, live classes at the weekend, and a capstone project you can show employers.