I turn messy data into models and dashboards people actually use.
Mostly Snowflake, dbt and Power BI, with machine learning where it helps.
English
- Model data in Snowflake and transform it with dbt - clean layers, tested models, KPIs that hold up.
- Build Power BI reports and dashboards, plus Streamlit apps when something more custom is needed.
- Handle the full pipeline: ingestion (SAP SuccessFactors, REST APIs, SFTP), ELT, CI/CD across environments, Terraform, Docker.
- Work on the ML side - forecasting, NLP, and putting models to work next to the data.
- Certified: Azure Data Engineer Associate (DP-203) · Fabric Analytics Engineer (DP-600)
- Preparing: SnowPro Core · dbt Analytics Engineering
- French (native) · English (C1) · Arabic (native) · Spanish (B1)
Data & analytics engineering
- fabric-movie-analytics - Bronze/Silver/Gold Medallion lakehouse, PySpark, OMDb enrichment, Power BI star schema. Runs locally with pandas too.
- data-lake-project - Snowflake data lake fed from GCS via Snowpipe, Streamlit insights app on top. Local DuckDB build included.
- london-boroughs-dataviz - scrape, clean, analyze, then an interactive Plotly/Streamlit dashboard with an OSM map.
Machine learning & AI
- imdb-rating-nlp - predict a film's rating from its plot. TF-IDF vs. sentence embeddings vs. DistilBERT. The simple baseline is hard to beat (MAE 0.67).
- socioeconomic-ml - Random Forest on 100k people, R² 0.68. Age and sex drive ~66% of the variance.
- movie-review-sentiment - scrape every review of a film, run transformer sentiment, roll it up into a reputation score.
Engineering & DevOps
- rest-api-nodejs - JWT auth, bcrypt, ownership checks, pagination, 9 Jest tests, Swagger docs, Docker.
- infrastructure-as-code - AWS two-tier web stack with Terraform, GitHub Actions CI, hardened security groups.
- iot-sensor-telemetry - ESP32 to MQTT to InfluxDB to Grafana, full stack runnable with no hardware.
aziz-ba.github.io/portfolio-website · LinkedIn · azizbenayed.pro@gmail.com
Version francaise
- Modelisation dans Snowflake avec dbt - couches propres, modeles testes, KPIs fiables.
- Rapports et dashboards Power BI, et applications Streamlit quand il faut quelque chose de plus personnalise.
- Toute la chaine : ingestion (SAP SuccessFactors, API REST, SFTP), ELT, CI/CD multi-environnements, Terraform, Docker.
- Cote ML : prevision, NLP, mise en production de modeles au plus pres de la donnee.
- Certifie : Azure Data Engineer Associate (DP-203) · Fabric Analytics Engineer (DP-600)
- En preparation : SnowPro Core · dbt Analytics Engineering
- Francais (natif) · Anglais (C1) · Arabe (natif) · Espagnol (B1)
Data & analytics engineering
- fabric-movie-analytics - Lakehouse Medallion Bronze/Silver/Gold, PySpark, enrichissement OMDb, schema etoile Power BI. Version pandas locale incluse.
- data-lake-project - Data lake Snowflake alimente depuis GCS via Snowpipe, application Streamlit. Build local DuckDB inclus.
- london-boroughs-dataviz - scraping, nettoyage, analyse, puis dashboard interactif Plotly/Streamlit avec carte OSM.
Machine learning & IA
- imdb-rating-nlp - Peut-on predire la note d'un film depuis son synopsis ? TF-IDF vs. embeddings vs. DistilBERT (MAE 0.67).
- socioeconomic-ml - Random Forest sur 100k personnes, R² 0.68. Age et sexe expliquent ~66% de la variance.
- movie-review-sentiment - scraping de critiques, analyse de sentiment par transformers, score de reputation global.
Ingenierie & DevOps
- rest-api-nodejs - Auth JWT, bcrypt, pagination, 9 tests Jest, Swagger, Docker.
- infrastructure-as-code - Stack AWS avec Terraform, CI GitHub Actions, groupes de securite renforces.
- iot-sensor-telemetry - ESP32 vers MQTT vers InfluxDB vers Grafana, stack complete sans materiel.
aziz-ba.github.io/portfolio-website · LinkedIn · azizbenayed.pro@gmail.com