Skip to content
View Salehmsa's full-sized avatar
4 followers · 24 following

Highlights

Block or report Salehmsa

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Salehmsa/README.md

Saleh Mahbub

Data Analyst · BI Developer · Data Scientist — Riyadh, Saudi Arabia

I build analytics systems end to end: raw source data → SQL Server star schema → DAX model → an executive report someone actually makes decisions from. Every KPI I publish is reconciled against the source database before it ships.

26 years inside the executive layer of a Saudi technology firm taught me the part most portfolios skip — what the business is actually asking before a single table is built.

Selected Work

Project What it is Stack
IT Sales & Marketing Intelligence B2B sales funnel for a Saudi system integrator. Star schema, RLS, bridge table for mixed-grain facts. 3,297 deals · SAR 193.8M won · 45.11% win rate — every figure reconciled to source SQL. Live dashboard → SQL Server · DAX · Power BI · Python
Flight Status Analysis 1.9M US domestic flights. Finding: 72.9% of delay minutes are airline-controllable, not weather. Ships a Python harness that asserts every published figure. Power BI · Python · pandas
NoorTrade — Saudi E-commerce 3-year performance audit, 8 cities, 27 KPIs. Ramadan +40%, White Friday +120%, Riyadh+Jeddah = 65% of revenue. Bilingual AR/EN dashboard. Python · pandas · Chart.js
Telco Customer Churn Churn prediction with hypothesis testing, XGBoost, and SHAP explainability — deployed as a Streamlit app. scikit-learn · XGBoost · SHAP · Streamlit
Salary Prediction Regression model with comparison across candidates, serialized for fast inference, served through an interactive app. scikit-learn · Streamlit
Africa Sales FY2022 9 dealer cities across Botswana & South Africa — trend, gap and anomaly analysis in a bilingual executive dashboard. Python · Chart.js

Built & Deployed

  • BI engineering — star schema design, USERELATIONSHIP for role-playing dates, bridge tables for mixed-grain facts, Row-Level Security, Time Intelligence
  • Machine learning — regression, classification, K-Means segmentation; XGBoost with SHAP explainability; model evaluation and feature engineering
  • Deployment — Streamlit and Gradio applications, serialized models, GitHub Pages dashboards
  • Data engineering — staging-table ETL, UTF-8 Arabic text handling, reproducible seeded data generation, KPI validation harnesses

Currently Going Deeper Into

RAG systems · FastAPI + Docker model serving · MLOps pipelines (CI/CD, experiment tracking) · Deep learning in PyTorch

Stack

Power BI DAX Power Query SQL Server T-SQL Python pandas NumPy scikit-learn XGBoost Matplotlib Seaborn Streamlit Git Docker FastAPI

Certified: IBM Data Science · IBM Data Analytics · Microsoft Power BI · PMP® · PMI-ACP®

📍 Riyadh · 📧 salehmahbub8@gmail.com · 💼 LinkedIn

Pinned Loading

  1. Saudi-Ecommerce-Analysis Saudi-Ecommerce-Analysis Public

    End-to-end sales analytics for a Saudi e-commerce company — 3 years, 27 KPIs, Python

    Jupyter Notebook 1

  2. Africa-Sales-Analytics-FY2022 Africa-Sales-Analytics-FY2022 Public

    End-to-end regional sales analytics · 9 dealer cities · Botswana & South Africa · FY 2022 · Python · Chart.js · Bilingual EN/AR Dashboard

    HTML

  3. Flight-Status-Data-Analysis Flight-Status-Data-Analysis Public

    End-to-end Power BI analytics on 1.9M US domestic flights (2015). 72.9% of delay minutes are airline-controllable, not weather. Ships a Python validation harness that asserts every published figure…

    Python

  4. IT-Sales-Marketing-Intelligence IT-Sales-Marketing-Intelligence Public

    End-to-end BI project: SQL Server → Star Schema → DAX → Power BI. Bilingual interactive dashboard included.

    HTML

  5. Salary-Prediction-Using-ML Salary-Prediction-Using-ML Public

    End-to-end Machine Learning project for salary prediction with model comparison, interactive Streamlit app, and data analysis.

    Jupyter Notebook

  6. Telco-Customer-Churn Telco-Customer-Churn Public

    End-to-end ML project: Predicting telecom customer churn with EDA, hypothesis testing, XGBoost, SHAP explainability, and Streamlit deployment.

    Jupyter Notebook