Hi, I'm
Tahar Guenfoud
Data Analyst & Data Scientist
A decade of IT education expertise combined with a Master's in Computer Science and intensive Data Science training. I bridge the gap between complex data and clear, impactful decision-making.
Featured Projects
Real-world data projects solving business problems
Blackspots — Railway Delay Cartography
Geospatial analysis of 19.7 million train records from Infrabel Open Data (Jan 2025 – Feb 2026). Interactive Folium heatmap identifying delay hotspots across 500 Belgian stations. Monthly trend analysis with Plotly, Streamlit dashboard, and optimized Parquet data pipeline.
Railway Delay ML — Predictive Modeling
End-to-end ML pipeline on 19.7M Infrabel Open Data records. Geographic clustering (DBSCAN haversine, 28 basins), anomaly detection (Isolation Forest, 50 stations), temporal decomposition (STL weekly), and multi-model comparison (XGBoost · GradientBoosting · RandomForest · Stacking). 5-Fold CV with sklearn Pipeline — no data leakage. AUC=0.813 · MAE=0.275 min (cross-validated).
Infrabel Dashboard — KPI Qualité Réseau
Dashboard interactif de ponctualité et fiabilité du réseau ferroviaire belge. 5 datasets Open Data Infrabel (ponctualité par gare, causes des retards, trains supprimés, Contrat de Performance). Notebook ETL/EDA complet + Streamlit + Power BI.
SNCB Live Dashboard
Real-time dashboard monitoring Belgian railway punctuality across 20 major stations. Live data from iRail API (parallel fetch via ThreadPoolExecutor), GPS coordinates extracted directly from API response. KPI gauges, Folium interactive map, Plotly heatmap, delay distribution, auto-refresh every 60s. Jupyter notebook with 12 analysis sections.
Bank Customer Churn Prediction
End-to-end ML pipeline to predict customer churn in banking. Identified 3 key factors impacting 40% of at-risk customers using behavioral analysis of 10,000+ users. Built predictive models (Random Forest, Logistic Regression) with advanced feature engineering.
Maven Music — Churn Analysis
Customer churn analysis for a music streaming platform. Reduction of customer attrition via behavioral analysis of 10,000+ users. Deployed interactive Streamlit dashboard accessible online.
Bank Customer Data Preparation
Comprehensive data preparation and cleaning pipeline for bank customer datasets. ETL processes, data quality assessment, and feature preparation for downstream ML models.
Le Wagon — Board Game Popularity Analysis
Capstone project for Le Wagon Data Science & AI bootcamp (400h+). Predictive modeling of board game popularity from 20,000+ BGG records enriched via REST API. Feature engineering, cross-validation, and Streamlit dashboard deployment.
Technical Skills
Languages & Data
Machine Learning
Visualization & BI
Cloud & Infrastructure
Experience & Education
Data Science & AI Bootcamp — 400h+
Le Wagon
Machine Learning, Deep Learning, Data Engineering, MLOps. Final team project with end-to-end deployment.
Master in Computer Science
UMONS — University of Mons
Advanced studies in algorithms, databases, and software engineering.
Computer Science Teacher
Fédération Wallonie-Bruxelles Enseignement · Mons
Teaching Python, SQL, and databases to 100+ students/year. Systematic analysis of performance data to adapt pedagogical strategies. Developed automated correction and reporting tools.
Independent IT Consultant
E-zzy · Belgium
Needs analysis, network architecture, and IT infrastructure deployment for SMEs. User training and technical support.
About Me
When I'm not crunching data, you'll find me training for my next triathlon 🏊🚴🏃. I believe that the discipline and analytical thinking required in sports directly translates to my approach in data science.