Skip to content
View Milad-Shabani's full-sized avatar
🏡
🏡

Block or report Milad-Shabani

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
Milad-Shabani/README.md
Milad Shabani - Business Intelligence Engineer

LinkedIn Instagram Email Website Resume Data interview question bank

🧭 About me

I turn complex business problems into systems, data and intelligence. From information systems and process analysis, through data engineering, to business intelligence and AI — I build solutions that make complexity understandable, data usable and decisions dependable.

Business Intelligence Engineer, 4+ years across telecom, pharmaceutical distribution, manufacturing and marketplace operations. M.Sc. in Industrial Engineering, Kharazmi University. Based in Karaj, Iran, and open to remote work and relocation.

Currently

  • Freelance BI engineer: ETL, warehouse design, Power BI dashboards, KPI frameworks
  • Publishing open analytics projects across steel trading, housing, finance, banking, marketplace, manufacturing, pharma and public data
  • Writing a Persian question bank for data interviews, free and growing

🏗️ Selected experience

AsiatechData warehouse serving 500M+ records, with the ETL and OLAP layers on top
Samanehaye ModiriatBI and CRM delivery for MAPNA, Digikala, MCI, Saba Battery
EltiamPharmC#/SSIS pipeline pulling from the APIs of 15 pharmaceutical distributors into one warehouse
Health Network MalardVaccine demand forecasting and vaccinator scheduling during the COVID-19 campaign

📂 Featured projects

Eighteen open-source projects, sixteen with a live dashboard, ordered by depth and quality. Start with these six:

Steel Trading: Where Do Deals and Days Leak? · live

A Microsoft Dynamics 365 export (Sales, Customer Service, shipments) from an iron and steel trader: 128K tons sold. Quotes sent within 4 hours win 74% of deals, against 37% after three days, and a win-probability model beats the reps' own CRM probabilities (AUC 0.67 vs 0.53). Lead to order, delivery, collections and support are modelled in BPMN 2.0 with measured numbers on every step: only 72% of orders arrive on time and in full, because loads wait 2.4 days in the yard for a truck.

Iran Housing: Where Does a Square Metre Cost What? · live · فارسی

516,947 real Divar listings and the Central Bank's transaction prices on hand-built maps of Tehran's 22 districts and Iran's 31 provinces. District 1 asks 4× District 17 per square metre, and the cheaper districts rose fastest. Set a budget to light up the neighbourhoods that fit, price a home with an in-browser model (R² 0.88), and test a 12-month forecast chosen by 35 historical backtests.

Supermarket Marketplace: Can Ops Survive a Mega-Sale? · live

578K orders across 5 fulfilment centres and 260 sellers, with mega-sales at 3–7x normal volume. A quantile forecast at 9.3% WAPE plans capacity around the events, and a churn model (AUC 0.78) tells the CRM team whom to win back.

Retail Banking: Which Transaction Is Suspicious? · live

Customer and transaction analytics with a suspicious-transaction model, and a dashboard that puts risk, customer behaviour and portfolio health side by side.

COVID Vaccination: Which Forecast Holds 45 Days Out? · live · فارسی

Rebuilt from my work at Malard Health Network: five centres, 500 to 3,500 doses a day. XGBoost matches SARIMA one step ahead, then collapses over 45 days without fresh actuals while SARIMA stays usable. The repo shows why.

Battery Plant: Can the CRM Pipeline Fix the Forecast? · live

Lumpy B2B orders leave Holt-Winters at 48% WAPE, so the probability-weighted CRM pipeline is blended in. Capacity is a linear program feeding MRP through a real BOM, and in the winter crunch it serves the higher-priced lines first.

Everything else

Project What it does Demo
Pharma Supply Chain: Why Is the Fill Rate Slipping? Fill rate fell from 84.5% to 81.0% while revenue grew: one LightGBM quantile model forecasts 588 SKU-centre series at 3.2% WAPE and sets the reorder points live
Unemployment Halved. Employment Didn't. Official unemployment fell from 12.4% to 7.5% while participation dropped 2.6 points: a bilingual data story with a model whose assumptions you can change live
Delivery Marketplace: Growth or Profit? Orders up 65% a year on a contribution margin under 7%: city P&L, cohorts, a 12–14% WAPE forecast and a reallocated promo budget live
Food Manufacturing S&OP Holt-Winters demand, capacity-constrained LP, MRP for a five-line plant live
Global Gold Market Pulse Real World Gold Council and IMF data: 26 years of reserves, production and demand, reconciled across sources live
Finance & FP&A A three-statement model that balances to the penny, a 12-month forecast and a budget-overrun risk model live
People Analytics An attrition-risk score with published weights across 740 employees and a 12-month hiring plan live
Resource & Capacity Planning A sales forecast turned into a day-by-day, DC-by-DC staffing plan live · فارسی
Global EV Market Pulse EV production, trade and batteries to 2035, on the IEA outlook live
Operations Analytics Throughput, service level and capacity across units live
User Analytics Pipeline Incremental event loads into a warehouse and a partitioned Parquet table code only
Power BI Daily Sales Dashboards shipped with the PBIX files, so the model and DAX are readable PBIX files

🧰 Tech stack

BI & reporting Power BI DAX Power Query SSRS Excel Tableau QlikView
Data & warehouse SQL Server T-SQL SSIS SSAS Parquet
Python & modeling Python pandas NumPy Matplotlib scikit-learn LightGBM and XGBoost statsmodels SciPy GAMS and CPLEX
Dashboards & delivery Chart.js Plotly D3 bpmn-js openpyxl GitHub Actions pytest
Enterprise & process Dynamics 365 Power Platform BPMN and UML Git Docker Jira

📚 Research & writing

  • M.Sc. thesis: a sustainable EV battery supply chain under uncertainty, with a data-mining approach to scenario generation
  • B.Sc. project: a review of data-mining methods for breast cancer prediction
  • Persian question bank for data interviews, plus teaching and Power BI training for teams

📊 Activity

Contribution streak Contribution graph

🤝 Let's work together

Open to collaboration: BI and analytics engineering roles, freelance warehouse and dashboard builds, and Power BI training for teams.

Email LinkedIn Website Resume

I work in Persian and English.

Pinned Loading

  1. Global-EV-market-pulse Global-EV-market-pulse Public

    World EV sales, production, batteries and the 2035 outlook, from the IEA Global EV Outlook 2026.

    JavaScript

  2. Food-manufacturing-sop-planning-engine Food-manufacturing-sop-planning-engine Public

    S&OP engine: turns a demand forecast into a feasible production plan (Holt-Winters, LP, MRP).

    Python 2

  3. Powerbi-analytics-online-market Powerbi-analytics-online-market Public

    Power BI dashboards and data models for online-market analytics, with PBIX files you can open.

    1

  4. Resource-planning-capacity-forecasting Resource-planning-capacity-forecasting Public

    Turns a sales forecast into a day-by-day, DC-by-DC staffing and delivery plan.

    Python 1

  5. Marketco-marketplace-intelligence Marketco-marketplace-intelligence Public

    Supermarket marketplace: 780K-row sales and CRM dataset with LightGBM forecasting and an RFM churn model.

    Python