TRIAL 2026-11-18
| Filename | Latest commit message | Latest commit date |
|---|---|---|
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X3rSFQDJYjX6bGw3aeA4iy |
||
| .gitignore | ||
| datalog1.csv | ||
| datalog2.csv | ||
| datalog3.csv | ||
| datalog4.csv | ||
| datalog5.csv | ||
| datalog_automatic1.csv | ||
| datalog_automatic2.csv | ||
| datalog_automatic3.csv | ||
| LICENSE | ||
| README.md | ||
subaru-datalogs
Raw COBB Accessport (AP3, v1.7.6) datalogs from a 2012 USDM Subaru Impreza WRX STI running a GT3076R turbo, ID1750x injectors, speed-density, flex-fuel base tune. Each CSV is one logging session at ~30 Hz: AF correction/learning, AFR, boost, load, timing, feedback/fine knock, DAM, injector duty, IAT, MAF, RPM, wastegate duty, etc. The last header column carries the AP/ROM info string.
datalog1..5.csv- manually triggered logsdatalog_automatic1..3.csv- auto-triggered logs
Tune (.ptm) files are not included (proprietary format). Load the CSVs with
anything, e.g. pandas.read_csv("datalog2.csv", encoding="latin-1") (degree
symbols are Latin-1), or the Datazap / COBB log viewers.
License: data released under MIT.