knoxel commited on
Commit
42507d4
·
verified ·
1 Parent(s): ebe0db1

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +27 -2
README.md CHANGED
@@ -30,6 +30,30 @@ python soccer_feature_engineering.py
30
 
31
  The script dynamically discovers all `*_dynamic_events.csv` files via `glob` — no hardcoded match IDs required.
32
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  ## Data Source
34
 
35
  [SkillCorner Open Data](https://github.com/SkillCorner/opendata) — 9 matches of Australian A-League 2024/2025, MIT License.
@@ -37,5 +61,6 @@ The script dynamically discovers all `*_dynamic_events.csv` files via `glob` —
37
  ## Requirements
38
 
39
  - Python 3.8+
40
- - pandas
41
- - numpy
 
 
30
 
31
  The script dynamically discovers all `*_dynamic_events.csv` files via `glob` — no hardcoded match IDs required.
32
 
33
+ ## Extended Analysis
34
+
35
+ `soccer_feature_engineering_extended.py` adds:
36
+
37
+ | Output | Description |
38
+ |--------|-------------|
39
+ | `features_first_half.csv` | All 33 features computed on 1st-half events only |
40
+ | `features_second_half.csv` | All 33 features computed on 2nd-half events only |
41
+ | `features_halves_diff.csv` | 2nd half minus 1st half — reveals tactical fatigue & in-game adjustments |
42
+ | `behavioral_fingerprint.csv` | 16-dim fingerprint per team-match (phase profile + pressing + runs + structure) |
43
+ | `cluster_labels.csv` | KMeans cluster assignments (k=4) on standardized fingerprint |
44
+ | `archetype_profiles.csv` | Centroid profiles for each discovered tactical archetype |
45
+ | `similarity_matrix.csv` | Pairwise cosine similarity between all 18 team-match fingerprints |
46
+ | `analysis.png` | Radar chart grid of the 4 archetypes |
47
+
48
+ ## Tactical Archetypes Discovered (k=4)
49
+
50
+ | Cluster | Signature | Characteristics |
51
+ |---------|-----------|-----------------|
52
+ | **0** — "Resilient Block" | Low build-up (71.6), modest pressing, balanced all-around | Conservative teams that stay compact; low risk, low reward |
53
+ | **1** — "Aggressive Dominator" | High build-up (85.0), massive pressing chains (231.7), elite off-ball runs (115.7) | High-press, high-tempo teams that generate territory and options |
54
+ | **2** — "Transition Specialist" | Low build-up (45.3), high quick_breaks (10.3), low runs (8.3) | Relies on direct play and transitions; minimal patient possession |
55
+ | **3** — "Organized Builder" | Highest build-up (90.3), balanced pressing, disciplined runs | Patient positional play with systematic defensive line pressure |
56
+
57
  ## Data Source
58
 
59
  [SkillCorner Open Data](https://github.com/SkillCorner/opendata) — 9 matches of Australian A-League 2024/2025, MIT License.
 
61
  ## Requirements
62
 
63
  - Python 3.8+
64
+ - pandas, numpy
65
+ - scikit-learn (for extended clustering)
66
+ - matplotlib, seaborn (for extended visualizations)