Hugging Face's logo

Datasets:
Mominullptr
/
fifa-world-cup-2026-dataset

Dataset card Data Studio Files
xet
 Community
Dataset Viewer
The dataset viewer is not available for this dataset.
Unexpected token '<', "<html> <h"... is not valid JSON

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.

FIFA World Cup 2026 Dataset- Live & Updated Stats — Matches, Squads, Players, Stats & Live Results

🔗 Dataset Links

The most complete and actively updated FIFA World Cup 2026 dataset on Kaggle. A clean, authentic, relational dataset covering the entire tournament (June 11 – July 19, 2026) with the first-ever 48-team format. Includes real match results updated daily, 1,248 players across all 48 squads, expected goals (xG), minute-by-minute match events, and per-team match statistics sourced from FIFA.com and verified providers.

Why this dataset? Unlike other World Cup datasets, this one is (1) updated daily with real match results as games are played, (2) fully relational with normalized foreign keys for SQL/database modeling, (3) contains zero synthetic data — every stat is sourced and traceable, and (4) includes advanced metrics like xG, match team stats, and VAR events.

🏆 Tournament Status

  • Tournament: FIFA World Cup 2026
  • Format: 48 Teams
  • Current Stage: Quarter-finals
  • Matches Completed: 96
  • Updated After: Every Completed Match
  • Validation: Relational integrity verified before every release

👥 Community Recognition

The FIFA World Cup 2026 Dataset has started gaining traction within the data science and developer community.

  • Mentioned by Software with Nick in an Instagram reel highlighting useful datasets for AI, machine learning, and data science projects.
  • Featured in the Analytics Jobs in Sports newsletter on Substack, highlighting the dataset for sports analytics.
  • Featured in a LinkedIn post that reached 22,000+ professionals, generating significant engagement from the data community.
  • Shared and discussed by developers building World Cup prediction models, dashboards, SQL projects, and sports analytics applications.
  • Continuously updated after every completed FIFA World Cup 2026 match.

Featured Mentions

Software with Nick (Instagram)

Analytics Jobs in Sports (Substack)

LinkedIn Community Post

Bissantz

Key Features

  • Real-World Group Configurations: Reflects the actual 12 groups (Groups A to L) with zero qualifiers placeholders.
  • Geographical & Altitude Details: Includes coordinates (Latitude/Longitude) and exact elevations in meters for all 16 host stadiums in the USA, Canada, and Mexico.
  • Granular Player Data: Features 1,248 players (26-man squads for all 48 teams) with market values in Euros, national team caps, positions, and club teams.
  • Live Expected Goals (xG): Includes expected goals metrics for all completed matches.
  • Match Events: Logs every goal, assist, card, and VAR review chronologically by minute.
  • Match Team Stats: Per-team per-match statistics (possession, shots, corners, etc.) sourced from verified providers.
  • Interactive Updater: Built-in interactive console application to easily record daily match outcomes and populate events without breaking database normalization.
  • ML-Ready Feature Set: A curated dataset with 65 pre-calculated features (ELO ratings, FIFA rankings, squad market values, rolling team form, and fatigue) ready to train machine learning classifiers.

🤖 ML-Ready Match Prediction Dataset

Skip the tedious feature engineering. This repository includes match_prediction_features.csv, a dataset engineered specifically for sports analytics and machine learning:

import pandas as pd
from xgboost import XGBClassifier

# Load dataset
df = pd.read_csv("match_prediction_features.csv")

# Train on completed matches (1 to 100), predict upcoming (101+)
train_df = df[df["match_id"] <= 100]
predict_df = df[df["match_id"] > 100]

# Features: Elo differences, Squad value ratio, Rolling form xG
features = ["home_elo", "away_elo", "home_fifa_rank", "away_fifa_rank", 
            "home_squad_total_value_eur", "home_prev_avg_xg_scored", "away_prev_avg_xg_scored"]

# Fit XGBoost Classifier to predict match result ('H', 'D', 'A')
clf = XGBClassifier()
clf.fit(train_df[features], train_df["match_result"])

# Predict upcoming knockout stages (Semi-Finals & Final)
predictions = clf.predict(predict_df[features])
print("Forecasted outcomes:", predictions)

Database Schema 

erDiagram
    TEAMS {
        int team_id PK
        string team_name
        string fifa_code
        string group_letter
        string confederation
        int fifa_ranking_pre_tournament
        int elo_rating
        string manager_name
    }
    VENUES {
        int venue_id PK
        string stadium_name
        string city
        string country
        int capacity
        float latitude
        float longitude
        int elevation_meters
    }
    TOURNAMENT_STAGES {
        int stage_id PK
        string stage_name
        bool is_knockout
    }
    REFEREES {
        int referee_id PK
        string name
        string country
        float avg_cards_per_game
    }
    MATCHES {
        int match_id PK
        string date
        string kickoff_time_utc
        int stage_id FK
        int venue_id FK
        int home_team_id FK
        int away_team_id FK
        int home_score
        int away_score
        string status
        float home_xg
        float away_xg
        int referee_id FK
    }
    SQUADS_AND_PLAYERS {
        int player_id PK
        int team_id FK
        string player_name
        string position
        string club_team
        int market_value_eur
        int caps
        string date_of_birth
        int height_cm
        int goals
    }
    MATCH_EVENTS {
        int event_id PK
        int match_id FK
        int minute
        string event_type
        int team_id FK
        int player_id FK
    }
    MATCH_LINEUPS {
        int lineup_id PK
        int match_id FK
        int player_id FK
        int team_id FK
        int is_starting_xi
        string tactical_position
        int minutes_played
    }
    PLAYER_STATS {
        int player_id PK, FK
        string player_name
        int team_id FK
        int matches_played
        int matches_started
        int minutes_played
        int goals
        int assists
        int yellow_cards
        int red_cards
        int penalty_scored
        int own_goals
        int saves
        int goals_conceded
        int clean_sheets
        string data_source
        string last_verified
    }

    TEAMS ||--o{ MATCHES : "hosts/visitors"
    TEAMS ||--o{ SQUADS_AND_PLAYERS : "roster"
    VENUES ||--o{ MATCHES : "hosts"
    TOURNAMENT_STAGES ||--o{ MATCHES : "stage"
    REFEREES ||--o{ MATCHES : "officiates"
    MATCHES ||--o{ MATCH_EVENTS : "contains"
    SQUADS_AND_PLAYERS ||--o{ MATCH_EVENTS : "triggers"
    MATCHES ||--o{ MATCH_LINEUPS : "lineups"
    SQUADS_AND_PLAYERS ||--o{ MATCH_LINEUPS : "plays_in"
    TEAMS ||--o{ MATCH_LINEUPS : "lineups"
    SQUADS_AND_PLAYERS ||--|| PLAYER_STATS : "has"
    TEAMS ||--o{ PLAYER_STATS : "has"
    MATCH_TEAM_STATS {
        int match_id FK
        int team_id FK
        int possession_pct
        int total_shots
        int shots_on_target
        int corners
        int fouls
        int offsides
        int saves
        string data_source
        string last_updated
    }
    MATCHES ||--o{ MATCH_TEAM_STATS : "stats"
    TEAMS ||--o{ MATCH_TEAM_STATS : "stats"

CSV Files Description

  1. teams.csv: Information on all 48 participating countries.
  2. venues.csv: Geolocation, capacities, and elevation details of all 16 stadiums.
  3. tournament_stages.csv: Lookup table for stages (Group Stage, Round of 32, etc.).
  4. referees.csv: International referees with their historical card-per-game stats.
  5. matches.csv: Match outcomes, dates, times, scores, xG metrics, and statuses using relational IDs (stage_id, venue_id, etc.) for clean database modeling.
  6. matches_detailed.csv: A denormalized, user-friendly version of matches.csv that displays human-readable names (e.g. home_team_name, stadium_name, city, referee_name) instead of IDs. Ideal for quick analysis without SQL joins!
  7. squads_and_players.csv: Detailed player registries (1,248 rows) containing verified player names (preserved with native accents), positions, clean club teams, market values, international caps, dates of birth (in YYYY-MM-DD format), heights in centimeters, and international goals.
  8. match_events.csv: Time-series game events (goals, assists, cards, VAR reviews) mapped to matches and players.
  9. match_team_stats.csv: Per-team per-match statistics (possession %, shots, shots on target, corners, fouls, offsides, saves) with data_source and last_updated columns for full traceability. Only populated with verified data from authentic sources (FIFA, Sofascore, FBref, etc.).
  10. match_lineups.csv: Tactical lineups for all completed matches: starting XI (11 players per team) and substitutes with actual minutes played.
  11. player_stats.csv: Cumulative tournament statistics for each player (1,248 rows), updated continuously as matches conclude. Outfield players have goalkeeper-specific fields set to NULL, and unverified advanced metrics (such as shots or key passes) are kept NULL to preserve dataset authenticity.
  12. match_prediction_features.csv: A machine-learning-ready features dataset containing 65 pre-calculated predictive features compiled for every match of the World Cup (completed and upcoming knockout stages). Features include team ratings (ELO, FIFA ranking, squad value), rolling team form (possession, shots, goals, xG, corners, saves), player fatigue (rest days), environmental factors (stadium elevation), and final target labels. Load directly into pandas or scikit-learn for rapid model training.

📋 Data Integrity Policy

Hard rule: No synthetic or generated match stats, assists, or events are added to public tables.

  • All match results, events, and statistics are sourced from verified, authentic providers.
  • New fields (e.g., assists, team stats) are only populated when confirmed from real sources.
  • The data_source column in match_team_stats.csv provides full traceability.
  • If verified data is unavailable for a match, that match is simply omitted from optional tables rather than filled with generated values.

🛠️ Installation & Usage

To generate the initial CSV dataset or regenerate it back to its default start state:

python generate_dataset.py

Daily Match Updates

As matches conclude every day, you can update the datasets interactively. The script will guide you step-by-step to record final scores, xG, and select players for goals, cards, and VAR reviews:

python update_dataset.py

SQLite Database Generation

For researchers and SQL query design, you can package all normalized CSV files into a single, query-optimized SQLite relational database file:

python generate_sqlite.py

🏷️ Citation

If you use this dataset in your research, publications, or projects, please cite it using the academic metadata provided in CITATION.cff or using the format below:

@dataset{fifa_world_cup_2026,
  author = {MD Mominul Islam},
  title = {FIFA World Cup 2026 Dataset- Live & Updated Stats},
  year = {2026},
  publisher = {Kaggle}
}

📄 License

This project is licensed under the Creative Commons Zero v1.0 Universal (CC0-1.0) Public Domain Dedication. Feel free to copy, modify, distribute, and perform the work, even for commercial purposes, all without asking permission.

🌐 Website

Explore the interactive database documentation, prediction guides, and sports analytics project ideas at: mominullptr.github.io/FIFA-World-Cup-2026-Dataset/
Downloads last month
517
Total file size:
1.86 MB