Skip to main content

PROJECT 02 • API / CLOUD / BI

Football Analytics Platform

An end-to-end football data platform that ingests API-Football data with Python, models it in Azure SQL and turns it into league, team and player analysis in Power BI.

PythonAzure SQLPower BIGitHub ActionsREST API

SCOPE

14 competitions

The reporting dataset covers the top three divisions in England, Spain, France and Germany, plus the top two divisions in Italy.

PIPELINE

API to reporting model

Python ingestion loads API responses into Azure SQL staging before SQL MERGE logic updates fact and dimension tables used by Power BI.

AUTOMATION

Scheduled ingestion

GitHub Actions runs the ingestion outside my local machine, with credentials and configuration managed through environment variables and secrets.

Architecture

API-Football → Python ingestion → Azure SQL staging → MERGE / upsert → Power BI

Separate ingestion modules handle leagues, teams, fixtures, standings, players, match statistics, events, lineups, injuries and transfers.

Engineering challenge

The API request allowance was not large enough to load every historical league, season and detailed endpoint in a single run. I handled this by batching historical ingestion by league and season while keeping the same pipeline and transformation logic.

Staging tables made it easier to inspect extracted data before updating the reporting model, while SQL upserts kept changing football data current without repeatedly inserting duplicates.

POWER BI DASHBOARD

From league overview to player comparison

The report moves from competition-level performance into individual team and player analysis. Select a dashboard view to open the screenshot at full size.

Open Power BI file ↗
Football Analytics Power BI overview dashboardOverviewStandings, scorers, assists, ratings and team metrics

Data model

The SQL model separates data by grain so league standings, matches and player performance can coexist cleanly in the same reporting model.

dim_competitiondim_teamdim_playerdim_datefact_matchfact_standingfact_player_seasonfact_player_matchfact_player_event

What I learned

This project gave me a much better understanding of how source-system limits, data grain, staging, incremental updates and automation affect the design of a real data pipeline rather than a one-off dashboard.