Thierry Donambi
$ pipelines

Pipelines

Selected data engineering projects: the ones with an ingest point, a processing stage, and something serving results at the end.

Spatial Selection Bias in Broadband Data

MSc thesis
Ookla + RURA data→ Observation model→ Cross-fit calibration→ Corrected estimates

Compares Ookla's self-selected speed tests against RURA's full subscriber census, then models the probability a location enters the crowdsourced sample to correct for it. Thirty quarters of measurements, terrain, and technology mix pulled from 28 regulator PDFs feed an out-of-fold calibration step.

  • Python
  • Geospatial
  • AWS Terrain Tiles
  • PDF Extraction
  • Propensity Modeling

Real-Time Fraud Detection

Source
Kafka→ Spark Structured Streaming→ Random Forest→ Postgres + HDFS

An end-to-end pipeline that simulates live transaction traffic and scores it in-flight. Predictions split two ways: Postgres for a low-latency Streamlit dashboard, HDFS as a data lake for batch analytics and retraining.

  • Python
  • Kafka
  • Spark Streaming
  • scikit-learn
  • PostgreSQL
  • HDFS
  • Streamlit

eCommerce Purchase Prediction

Source
HDFS→ Kafka (WebHDFS)→ Spark Streaming (MLlib)→ MySQL→ Django

A fully containerized clickstream pipeline that predicts whether a shopping session will end in a purchase while the session is still in flight, surfaced on a live dashboard. Built with a five-person team for AUCA's Big Data Essentials course.

  • Docker Compose
  • HDFS
  • Kafka
  • PySpark MLlib
  • MySQL
  • Django

Smart Energy Grid Analytics

Source
Sensors (MQTT / EMQX)→ TimescaleDB→ Grafana + Streamlit

Ingestion and analytics pipeline for smart-grid sensor data. Devices publish over MQTT through EMQX into a Postgres/TimescaleDB store, surfaced through Grafana boards and an interactive Streamlit dashboard.

  • Python
  • MQTT
  • EMQX
  • TimescaleDB
  • PostgreSQL
  • Grafana

Hadoop vs. Spark, Head-to-Head

Source
Hadoop MapReduce vs PySpark

A Master's-level benchmark comparing MapReduce against Spark on the IEEE-CIS fraud dataset (590,540 transactions, 394 columns), measuring performance and implementation complexity, not just output.

  • Hadoop
  • PySpark
  • Schema-on-Read
$ applied-systems

Applied systems

Where the data work meets a real product: deployed for actual users, not just a notebook.

Mr. Cocorico

Butchery management and e-commerce platform running in production on Railway, with inventory tracking and POS logic for a real business.

  • Django
  • PostgreSQL
  • Railway

PeaceData Portal

Data and project-management platform for peacebuilding initiatives, with role-based access separating project managers, field data engineers, analysts, and administrators.

  • Django
  • RBAC
  • Dashboards

Smart Greenhouse Monitoring

IoT dashboard on AWS RDS with live Chart.js visualization and automated email alerts when sensor thresholds are breached.

  • Django REST
  • AWS RDS
  • Chart.js

Hope for Kids Chad

Bilingual English/French platform built for an NGO's regional expansion.

  • React
  • TypeScript

VetConnect Rwanda

Django platform connecting smallholder livestock farmers with licensed veterinarians: AI-assisted symptom triage, mobile-money consultations, and digital prescriptions. Piloting in Nyagatare District.

  • Django
  • Mobile Money
  • Gemini AI
$ more-projects

More projects

  • colouring-book-press

    Turns a topic into a print-ready paperback: plans the subject list, generates fifty illustrations, vector-traces them, and lays out interior and cover for both French and English editions.

  • unmapped

    AI pipeline turning informal skills into structured, ESCO-aligned profiles, built for the World Bank Youth Summit × Hack-Nation challenge.

  • association-rule-learning-groceries

    Apriori-based market basket analysis over 9,835 grocery transactions.

  • investmatch-hackathon

    Rule-based Streamlit tool recommender for forensic investigators: no AI, no API costs, just scored matches with explanations.

  • neo4j-project

    Graph database modeling and querying for a coursework assignment.

  • Netflix-User-Activity-Analysis

    Viewing-behavior insights engineered from raw watch history.

  • Spotify-Streaming-History

    Listening-history analysis with pandas, NumPy, and matplotlib/seaborn.

  • Cafe-Sales-kaggle

    Cleaning a Kaggle sales dataset and loading it into Postgres.

  • domain-lookup-tool

    CLI and GUI utility for domain name lookups.