How r-niemans/vds_visuals2425 works
Project type and purpose
• This is a data-visualization web application built in Python.
• It consumes raw football match and player data (CSV) to generate static and interactive charts, then serves them via a simple web interface (app.py).
File structure
• Top-level scripts and entry point:
• app.py – the Flask (or similar) server that ties everything together
• Jupyter notebook (VDS2425_Football.ipynb) – for exploratory analysis
• plot-generation scripts (bar_and_violin_plot.py, heatmap.py, scatterplot.py, etc.)
• Data directories:
• data/ – raw source CSVs for Country, League, Match, Player, Team, etc.
• grouped_players/ and filtered CSVs – intermediate data derived from raw inputs
• Output assets:
• figures/ – JSON or serialized plot objects for reuse in the app
• assets/ – static files (CSS, JS, images) used by the front end
• Configuration / metadata:
• README.md – minimal, project name only
• .idea/ – IDE settings (can be ignored for architecture)
Extract insights from the README
• Project name “vds_visuals2425” implies “visualizations for VDS 2425 (a football dataset).”
Main components
a. Data Layer
• Raw CSV files in data/
• Preprocessing scripts that read raw data and write grouped_players/ and filtered CSVs
b. Processing & Visualization Layer
• Standalone Python scripts and notebooks that generate:
– Static charts (bar_and_violin_plot.py, heatmap.py, scatterplot.py)
– Interactive charts (bar_and_violin_interactive.py)
• Output stored as JSON or Python objects in figures/
c. Web Application Layer
• app.py – reads precomputed figures or regenerates plots on demand
• assets/ – static front-end resources (JS libraries for interactivity, CSS)
• Templates or HTML endpoints (implied inside app.py)
d. Client Layer
• Browser – requests pages, loads interactive charts via JS
Architectural patterns and principles
• Modular separation:
• Data extraction/prep scripts isolated from web-serving logic
• Visualization logic separated into self-contained scripts
• Reusability: pre-generated figures live in figures/ so the web app need not rerun heavy plotting on every request
• Single codebase monolith: no microservices, everything runs under one Python process
Technologies, frameworks, and libraries
• Python 3.x
• Flask (or lightweight HTTP server within app.py)
• Plotting libraries: Matplotlib/Seaborn for static, Plotly for interactive
• Pandas for CSV ingestion and DataFrame manipulation
• JSON for serializing interactive figures