π I'm a passionate Data Scientist who loves building things and solving real peopleβs problems using data.
π I turn raw data into insights, automate painful processes, and create tools that make peopleβs lives easier.
π¨βπ» Currently working on data science, machine learning, and end-to-end data solutions (from extraction to visualization and deployment).
π€ Always open to collaborating on impactful data projects, automation tools, or anything that solves a real problem.
π± Constantly learning and deepening my skills in ML/AI, cloud data platforms, and advanced analytics.
π¬ Ask me about: Python, Data Science, Machine Learning, SQL, Data Engineering, Automation, Web Scraping, APIs, Pandas, NumPy, Scikit-learn, TensorFlow, PyTorch, Data Visualization (Matplotlib, Seaborn, Plotly), ETL pipelines, Linux scripting, and more.
①Fun fact: I learned C and Java before HTML/CSS⦠and then discovered the magic of turning data into solutions people actually use!
Highlights
- Pro
Pinned Loading
-
Data-analysis
Data-analysis PublicData analysis portfolio across Python, R, SQL and Power BI β retail sales, customer ,segmentation, Naive Bayes, PCA, HR absenteeism and BI dashboards.
Jupyter Notebook
-
Exploratory-Data-Analysis
Exploratory-Data-Analysis PublicTwo EDA studies in pandas and seaborn β the Roller Coaster Database and world population 1970β2022 β each going from raw CSV to cleaned data to a closing question.
Jupyter Notebook
-
loan-eligibility-prediction
loan-eligibility-prediction PublicLoan approval classification with EDA, missing-value treatment and log-skew correction. Naive Bayes reaches 0.83 accuracy against a 0.69 majority-class baseline.
Jupyter Notebook
-
morocco-banks-reviews
morocco-banks-reviews PublicEnd-to-end analytics pipeline over Google Maps reviews of Moroccan bank branches β Scrapy/Playwright scraping, dbt cleaning, French NLP enrichment, and a star schema for Looker Studio.
Python
-
ppe-detection
ppe-detection PublicConstruction-site PPE detection using a fine-tuned YOLOv8 model served through a Flask app. 14 classes covering both worn and missing equipment.
Python
-
vision-anomaly-detection
vision-anomaly-detection Publicβ Four unsupervised anomaly detection methods benchmarked on MVTec AD (Pill) β autoencoder, deep feature reconstruction, global KNN and PatchCore-style patch KNN. Best AUROC 0.943.
Jupyter Notebook
If the problem persists, check the GitHub status page or contact support.