This is where the detailed record of every team session lands. The *.csv
files are not tracked in the repository (they are in .gitignore) — they are
the local output of the game, copied off after the event for analysis in
Pandas/Excel.
| File | Granularity | Columns |
|---|---|---|
../leaderboard.json |
top 10 across sessions | team, score, grade, ts |
../summary.csv |
1 line / session | session_id, team, timestamp, total_score, grade, rank |
sessions/<id>_<team>.csv |
1 line / joint / player / round | below |
Full column set of a per-session file:
session_id, team, timestamp, round_idx, pose, difficulty, multiplier,
player_idx, player_score, joint, target_angle, achieved_angle,
joint_score, visible, weight
import pandas as pd
df = pd.read_csv("sessions/2026-05-21_14-05-33_feniksiarze.csv")
# Mean score per joint — which joint cost the most points?
df.groupby("joint")["joint_score"].mean().sort_values()
# Hardest poses across the whole event
import glob
allf = pd.concat(pd.read_csv(f) for f in glob.glob("sessions/*.csv"))
allf.groupby("pose")["joint_score"].mean().sort_values()