Market data is never clean. Can Muse catch it?
Muse Spark reads a CSV of market data and flags duplicates, gaps, bad prices and unadjusted splits. A deterministic scorer checks its answer. No sign-in required.
Try a sample:
or
Results
Sample data
Download sample CSVs
Each sample contains equity daily bars with planted data-quality problems. Download one, upload it above, and see what the model catches.
task_001.csv
- Duplicate row (2024-04-10)
- Fat-finger price spike (high at 1096.5)
- Stock split not back-adjusted
task_002.csv
- Duplicate row (2024-07-18)
- Missing day (2024-07-15, weekday)
- Stock split not back-adjusted
task_003.csv
- Missing day (2024-04-30, weekday)
- Bad price spike (close at 1065.83)
- Stock split not back-adjusted
Build with LaneMint
Quant-engineering tasks with graders that check the details. Deterministic, sandboxed, runs on hidden synthetic data.