Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 12 new columns ({'away_team', 'league', 'home_team', 'pick_goals', 'ft_score', 'date', 'pick_ou', 'pick_all', 'pick_htft', 'ht_score', 'pick_1x2', 'pick_score'}) and 31 missing columns ({'yellow_cards', 'accurate_crosses_per_match', 'shots_on_target_per_match', 'penalties_conceded', 'xg_diff', 'interceptions_per_match', 'clean_sheets', 'goals_per_match', 'accurate_long_balls_per_match', 'xg', 'counterpress_per_match', 'tackles_per_match', 'xga', 'corners', 'set_piece_goals_conceded', 'penalties_won', 'saves_per_match', 'team', 'avg_rating', 'set_piece_goals', 'big_chances_missed', 'red_cards', 'distance_per_match', 'big_chances', 'fouls_per_match', 'touches_in_opp_box_per_match', 'possession_pct', 'goals_conceded_per_match', 'accurate_passes_per_match', 'clearances_per_match', 'avg_attendance'}).

This happened while the csv dataset builder was generating data using

hf://datasets/xgmind/premier-league-xg-audit/xgmind_prediction_ledger.csv (at revision 2b73a8d25e84d5330967d36a437ba1d688446415), ['hf://datasets/xgmind/premier-league-xg-audit@2b73a8d25e84d5330967d36a437ba1d688446415/premier_league_xg_audit_2025_26.csv', 'hf://datasets/xgmind/premier-league-xg-audit@2b73a8d25e84d5330967d36a437ba1d688446415/xgmind_prediction_ledger.csv']

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1848, in _prepare_split_single
                  writer.write_table(table)
                  ~~~~~~~~~~~~~~~~~~^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 765, in write_table
                  self._write_table(pa_table, writer_batch_size=writer_batch_size)
                  ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              date: string
              home_team: string
              away_team: string
              league: string
              ft_score: string
              ht_score: string
              pick_1x2: string
              pick_goals: string
              pick_score: string
              pick_htft: string
              pick_all: string
              pick_ou: string
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 1670
              to
              {'team': Value('string'), 'avg_rating': Value('float64'), 'xg': Value('float64'), 'xga': Value('float64'), 'xg_diff': Value('float64'), 'goals_per_match': Value('float64'), 'goals_conceded_per_match': Value('float64'), 'possession_pct': Value('float64'), 'big_chances': Value('int64'), 'big_chances_missed': Value('int64'), 'shots_on_target_per_match': Value('float64'), 'accurate_passes_per_match': Value('float64'), 'accurate_long_balls_per_match': Value('float64'), 'accurate_crosses_per_match': Value('float64'), 'touches_in_opp_box_per_match': Value('int64'), 'corners': Value('int64'), 'set_piece_goals': Value('int64'), 'set_piece_goals_conceded': Value('int64'), 'penalties_won': Value('float64'), 'penalties_conceded': Value('float64'), 'clean_sheets': Value('int64'), 'saves_per_match': Value('float64'), 'tackles_per_match': Value('float64'), 'interceptions_per_match': Value('float64'), 'clearances_per_match': Value('float64'), 'counterpress_per_match': Value('float64'), 'fouls_per_match': Value('float64'), 'yellow_cards': Value('int64'), 'red_cards': Value('float64'), 'distance_per_match': Value('float64'), 'avg_attendance': Value('int64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1850, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
                  ...<4 lines>...
                  )
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 12 new columns ({'away_team', 'league', 'home_team', 'pick_goals', 'ft_score', 'date', 'pick_ou', 'pick_all', 'pick_htft', 'ht_score', 'pick_1x2', 'pick_score'}) and 31 missing columns ({'yellow_cards', 'accurate_crosses_per_match', 'shots_on_target_per_match', 'penalties_conceded', 'xg_diff', 'interceptions_per_match', 'clean_sheets', 'goals_per_match', 'accurate_long_balls_per_match', 'xg', 'counterpress_per_match', 'tackles_per_match', 'xga', 'corners', 'set_piece_goals_conceded', 'penalties_won', 'saves_per_match', 'team', 'avg_rating', 'set_piece_goals', 'big_chances_missed', 'red_cards', 'distance_per_match', 'big_chances', 'fouls_per_match', 'touches_in_opp_box_per_match', 'possession_pct', 'goals_conceded_per_match', 'accurate_passes_per_match', 'clearances_per_match', 'avg_attendance'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/xgmind/premier-league-xg-audit/xgmind_prediction_ledger.csv (at revision 2b73a8d25e84d5330967d36a437ba1d688446415), ['hf://datasets/xgmind/premier-league-xg-audit@2b73a8d25e84d5330967d36a437ba1d688446415/premier_league_xg_audit_2025_26.csv', 'hf://datasets/xgmind/premier-league-xg-audit@2b73a8d25e84d5330967d36a437ba1d688446415/xgmind_prediction_ledger.csv']
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

team
string
avg_rating
float64
xg
float64
xga
float64
xg_diff
float64
goals_per_match
float64
goals_conceded_per_match
float64
possession_pct
float64
big_chances
int64
big_chances_missed
int64
shots_on_target_per_match
float64
accurate_passes_per_match
float64
accurate_long_balls_per_match
float64
accurate_crosses_per_match
float64
touches_in_opp_box_per_match
int64
corners
int64
set_piece_goals
int64
set_piece_goals_conceded
int64
penalties_won
float64
penalties_conceded
float64
clean_sheets
int64
saves_per_match
float64
tackles_per_match
float64
interceptions_per_match
float64
clearances_per_match
float64
counterpress_per_match
float64
fouls_per_match
float64
yellow_cards
int64
red_cards
float64
distance_per_match
float64
avg_attendance
int64
Arsenal
7.06
65.6
28.3
37.3
1.9
0.7
56.2
116
78
4.9
396.7
18.4
4.3
1,255
216
23
7
4
null
19
1.6
15.8
7.1
23.8
4.7
10.3
51
null
117,707.8
60,217
Manchester City
7.22
71.6
44.2
27.4
2
0.9
60.5
120
74
5.4
520.4
19.4
4.4
1,409
244
11
12
4
2
16
2.4
14.4
7.9
21.6
4.6
9.7
67
null
118,540.1
52,730
Manchester United
7.08
65.3
48.6
16.8
1.8
1.3
51.8
109
68
5.7
385.8
22.8
4.6
1,047
181
19
14
6
4
8
2.4
18.2
8.6
25
3.4
10.3
62
3
113,036.6
73,975
Chelsea
6.89
66.5
52.3
14.3
1.5
1.4
57.7
102
57
4.4
471.2
20.4
3.7
1,005
226
13
17
8
2
9
2.8
15.6
9.8
21.9
3.6
11.1
90
8
109,744.9
39,602
Liverpool
7
61.3
47.4
13.9
1.7
1.4
59.3
100
64
4.6
455.4
21
4.4
1,278
232
18
20
2
4
10
2.6
13
6.9
25.2
4.1
10.1
57
1
111,699.7
60,389
Brighton & Hove Albion
6.88
57.9
49.1
8.8
1.4
1.2
54.1
88
56
4.6
400.3
20.4
5.1
1,074
185
11
7
5
6
10
2.8
17.8
8.5
23.6
5.1
11.9
86
null
117,167.4
31,379
Crystal Palace
6.84
58.2
51.9
6.3
1.1
1.3
45.9
98
66
3.7
306.9
20
3.7
905
159
10
18
8
7
12
2.9
18.6
8.3
33.7
3.1
10.5
74
2
115,580.8
25,017
Newcastle United
6.85
57.1
51.3
5.8
1.4
1.4
52.6
90
59
4.6
374.3
20.7
5
990
230
15
17
6
5
8
3
14.5
7.3
26.4
3.7
10.2
63
3
117,375.5
52,161
Brentford
6.85
60.3
54.4
5.8
1.4
1.4
47.6
114
69
3.9
313.2
23.4
4.8
944
182
10
8
10
8
10
2.9
14.6
8.7
29.2
3.1
10.3
69
1
116,253.6
17,132
AFC Bournemouth
6.94
61.9
56.8
5.1
1.5
1.4
50.1
89
54
4.7
331.4
20
3.5
950
212
14
18
6
6
11
2.9
17.2
9.6
30.9
4.6
12
88
2
113,919.4
11,175
Leeds United
6.82
54.2
55.7
-1.4
1.3
1.5
45.8
89
57
4
314.1
20.6
4.8
899
168
16
17
8
6
8
2.7
17.4
8.9
30.5
4.1
10.5
62
1
118,042.1
36,695
Fulham
6.81
48.4
53.4
-5.1
1.2
1.3
51.7
61
37
3.7
397.8
21.7
4.6
951
187
11
9
5
4
9
2.6
17.1
8.6
26.6
2.5
10.9
75
1
115,223
27,228
Aston Villa
6.81
48.2
53.9
-5.7
1.5
1.3
53.6
87
61
4.5
384.6
19.8
3.6
981
199
18
15
null
2
9
3
15.4
6.3
21
3.4
10
58
1
112,445.2
41,977
Everton
6.89
46.7
56.5
-9.8
1.2
1.3
43.7
71
42
3.7
303.6
21.4
4.1
894
168
13
11
2
4
11
2.6
18.5
8.2
29.8
4
10.8
72
4
112,065.9
52,132
Nottingham Forest
6.82
46
57.3
-11.3
1.3
1.3
46.8
69
39
4.1
335.2
21.8
4.6
844
198
11
13
3
6
9
2.9
16.9
7.9
25.6
3.8
10.7
60
1
111,796.9
30,425
Tottenham Hotspur
6.78
40
52.3
-12.3
1.3
1.5
50.4
71
45
3.8
337.6
20
4.9
924
209
19
10
null
3
9
2.4
19.6
8.4
26.6
3.9
12.2
99
4
112,599.5
61,006
West Ham United
6.78
46.1
59.6
-13.5
1.2
1.7
42.5
72
46
3.7
289.9
20.2
3.7
771
187
14
17
3
7
7
3.5
17.7
9.1
29
3.3
11.1
68
3
112,498.7
62,347
Sunderland
6.79
38.9
54.4
-15.6
1.1
1.3
44.9
64
42
3.4
297.5
20
3.1
764
138
9
14
5
3
11
3.1
15.4
7.9
30.5
3.1
10.7
82
3
112,617.6
46,546
Wolverhampton Wanderers
6.59
35.5
59.5
-24
0.7
1.8
42.7
54
36
3.2
289
21.1
5.2
756
129
7
11
5
5
4
2.9
19.1
8.5
28.9
2.6
13
77
3
112,704
29,835
Burnley
6.57
32.5
75.4
-42.8
1
2
42.3
55
33
3.1
304.4
21.3
2.9
650
138
7
14
2
8
4
3.6
16.8
9
30.1
2.8
10.4
64
3
116,390.4
21,005
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null

XG Mind AI — Public Football Prediction Ledger & Premier League xG Audit

Two open datasets from XG Mind AI (https://xgmind.com/), an AI-driven football match-intelligence site that publishes every prediction and settles it in public — losses included, unedited. Live ledger: https://xgmind.com/performance/ · Methodology: https://xgmind.com/methodology/

Dataset Description

xgmind_prediction_ledger.csv — 37 settled matches (2026-09-29 → 2026-10-01, snapshot 2026-10-01). A row-for-row transcription of the "Match-by-match ledger" at https://xgmind.com/performance/. Each row is one match report with the published settlement per dimension (HIT/MISS):

  • 1X2 31/37 (83.8%) — full-time result matches the published direction
  • Goals 20/37 (54.1%) — actual total goals is one of the predicted goal counts
  • Score 15/37 (40.5%) — final score is one of the predicted scorelines
  • HT/FT 27/37 (73.0%) — half-time/full-time combination matches
  • All 9/37 (24.3%) — all four dimensions above hit
  • O/U 17/37 (45.9%) — the 2.5-goal line, settled independently

premier_league_xg_audit_2025_26.csv — season-level advanced metrics for all 20 Premier League clubs, 2025-26 season (38 rounds): xG, xGA, xG difference, possession %, big chances created/missed, shots on target, passing accuracy splits, touches in the opposition box, corners, set-piece goals for/against, penalties won/conceded, clean sheets, saves, tackles, interceptions, clearances, counter-pressing actions, fouls, cards, distance covered and average attendance. Genuine gaps are left empty (NaN) — never zero-filled — so NaN-aware statistics stay honest (cf. https://xgmind.com/blog/missing-football-data-is-not-zero/).

Data Fields

xgmind_prediction_ledger.csv

column type description
date string (YYYY-MM-DD) match date
home_team string home team
away_team string away team
league string competition
ft_score string full-time score, "home-away"
ht_score string half-time score, "home-away" (empty when not published)
pick_1x2 HIT/MISS 1X2 direction settlement
pick_goals HIT/MISS total-goals settlement
pick_score HIT/MISS correct-score settlement
pick_htft HIT/MISS half-time/full-time settlement
pick_all HIT/MISS all-four-dimensions settlement
pick_ou HIT/MISS over/under 2.5 settlement

premier_league_xg_audit_2025_26.csv

column type description
team string club name
avg_rating float average match rating
xg float expected goals, season total
xga float expected goals against, season total
xg_diff float xG − xGA
goals_per_match float average goals scored
goals_conceded_per_match float average goals conceded
possession_pct float average possession %
big_chances int big chances created
big_chances_missed int big chances missed
shots_on_target_per_match float average shots on target
accurate_passes_per_match float average accurate passes
accurate_long_balls_per_match float average accurate long balls
accurate_crosses_per_match float average accurate crosses
touches_in_opp_box_per_match float average touches in opposition box
corners int corners, season total
set_piece_goals int goals from set pieces
set_piece_goals_conceded int goals conceded from set pieces
penalties_won float penalties won (NaN where not recorded)
penalties_conceded float penalties conceded (NaN where not recorded)
clean_sheets int clean sheets
saves_per_match float average saves
tackles_per_match float average tackles
interceptions_per_match float average interceptions
clearances_per_match float average clearances
counterpress_per_match float average counter-pressing actions
fouls_per_match float average fouls committed
yellow_cards int yellow cards
red_cards float red cards (NaN where not recorded)
distance_per_match float distance covered — transcribed verbatim from the source column labeled "km"; magnitudes (~117,708 for Arsenal) suggest meters, reproduced as-is
avg_attendance int average home attendance

Data Splits

Single split per file (train-style full tables): the ledger has 37 rows, the xG audit has 20 rows. No train/test split — these are reference/audit tables.

Usage

import pandas as pd

ledger = pd.read_csv("xgmind_prediction_ledger.csv")
hit_rates = (ledger.filter(like="pick_") == "HIT").mean().round(3)
print(hit_rates)
# pick_1x2     0.838
# pick_goals   0.541
# pick_score   0.405
# pick_htft    0.730
# pick_all     0.243
# pick_ou      0.459

xg = pd.read_csv("premier_league_xg_audit_2025_26.csv")
print(xg[["team", "xg", "xga", "xg_diff"]].sort_values("xg_diff", ascending=False).head())
# NaN-aware by default: xg["penalties_won"].mean() skips missing values (correct);
# xg["penalties_won"].fillna(0).mean() would bias the mean downward (wrong).

Citation

@dataset{xgmind_ledger_xg_audit_2026,
  title  = {XG Mind AI --- Public Football Prediction Ledger \& Premier League xG Audit},
  author = {{XG Mind AI}},
  year   = {2026},
  url    = {https://xgmind.com/performance/}
}

License

CC-BY 4.0 — share and adapt with attribution to "XG Mind AI (https://xgmind.com/)". Contact: press@xgmind.com.

Downloads last month
27