In this Repo
This project fetches data for 2024-2025 season from FBref and creates an scatter plot visualization showing which teams are overperforming or underperforming relative to their Expected Goals (xG).
Pressure events, percentage point difference from Serie A 2021/22 after Match Day 20
Parsing raw StatsBomb data and storing it in a Pandas DataFrame
This script transforms a player's FBref URL into 9 separate CSV files containing different Stats categories.
Uses the Exa Agent API to produce tactical and career research briefs for football managers. Exa Agent combines frontier language models with Exa's web search tools to deliver structured, source-backed research through a single API call.
The notebook (EXA_Agent_Extended_Football_Managers.ipynb) researches each manager across six dimensions:
- Tactical philosophy & preferred system
- Career history & trophy haul
- Key strengths
- Known weaknesses & criticisms
- Peer comparisons & rivalries
- Source URLs
Edit the MANAGERS list to research any set of managers:
MANAGERS = [
{"manager": "Jose Mourinho"},
{"manager": "Pep Guardiola"},
]- An Exa API key stored in Google Colab secrets as
EXA_KEY pip install exa_py
| Effort | Cost | Best for |
|---|---|---|
minimal |
$0.012 / request | Lightweight tasks, lowest cost |
low |
$0.025 / request | Simple lookups, narrow factual tasks |
medium |
$0.10 / request | Standard research tasks |
high |
$0.50 / request | Harder research, more citations |
xhigh |
$1.00 / request | High-value tasks where completeness matters |
The notebook uses auto effort, which dynamically scales compute to the task.
Links: Exa Agent Blog Post · API Docs · Playground