Skip to content

Repository files navigation

Data Science Portfolio: Technical Projects

🌐 Language / Idioma: English Version 🇺🇸 | Versão em Português 🇧🇷

Python Machine Learning Google BigQuery

Author: Lucas Nascimento Oliveira
Role: Data Scientist


📍 Strategic Navigation Guide

Click on the links below to explore projects by seniority level, technical complexity, and generated business ROI.

Level Project Technical Focus Methodology & Business ROI Documentation
Advanced Financial Fraud Analytics Cloud MPP (BigQuery) & SQL Distributed query optimization for cost reduction (OPEX) in corporate Data Lake. Case Study
Advanced BigQuery LTV Prediction BQML & Advanced SQL (Cohort/RFV) Native linear regression modeling in Data Warehouse to forecast future revenue by customer cohorts. Case Study
Advanced Aviation Ops Risk Random Forest & SHAP (XAI) Mitigating $960k USD (R$ 4.8M) in operational risks in aviation networks using explainable AI. Case Study
Advanced Hospital Risk Audit Isolation Forest (Outliers) Automated identification of R$ 2.4M ($480k USD) in medical billing anomalies and hospital claims. Case Study
Intermediate Customer Segmentation K-Means Clustering RFV behavioral segmentation for customer acquisition and retention cost optimization in marketing campaigns. Case Study
Intermediate Market Basket Analysis Association Rules & Bundling Apriori algorithm applied to transactions for average ticket optimization through product bundling. Case Study
Foundational Pricing Intelligence Big Data Viz & PCI Index Calculation of the Price Competitiveness Index (PCI) on 370k+ daily competitor pricing records. Case Study
Foundational Geomarketing Expansion Geospatial Density Analytics Density mapping of UK EV charging infrastructure to optimize expansion CAPEX. Case Study

Tip

Master Access: For a consolidated view of the entire technical journey, please refer to walkthrough_master.md.



📈 Financial Impact & Methodology (Simulated Cases)

Note

The financial projections below represent estimates and study cases for business validation under simulated scenarios, showing how data science algorithms drive EBITDA and reduce OPEX.

Estimated Impact in Simulated Scenario: This portfolio demonstrates the practical application of data science techniques focused on business process optimization and risk mitigation, estimated at up to R$ 7.2M+ (approx. $1.4M+ USD) of projected value under simulated scenarios (case studies), with potential OPEX reduction of up to 15% and simulated predictive accuracy of 82% in critical decisions.

  • Impact Methodology (R$ 7.2M+ / $1.4M+ USD - Simulated Scenario Estimate):
    • R$ 4.8M ($960k USD): Projected annual revenue protected under simulated aviation network scenario via the predictive mitigation model for severe delays (Aviation Ops Risk).
    • R$ 2.4M ($480k USD): Estimated savings in simulated medical billing audit and entry errors identified by anomaly detection AI (Hospital Risk Audit).
graph TD
    A["Data Ecosystem: Projected Impact $1.4M+ USD (Simulated Scenario)"] --> B["Estimated Logistic Risk Mitigation $960k USD"]
    A --> C["Estimated Health Billing Protection $480k USD"]
    B --> B1[Delay Prediction via Random Forest]
    B --> B2[Explainability of Root Causes via SHAP]
    C --> C1[Audit/Double-Billing Detection via AI]
    C --> C2[Automated Auditing via Isolation Forest]
Loading

🏗️ Project Architecture (Modularity)

The project follows a professional structure geared towards scalability and production-grade standards:

  • /data/processed: Cleaned and scored datasets ready for consumption.
  • /src/ml: Machine Learning engine (model_engine.py).
  • /src/reporting: Corporate visualization module (viz_factory.py).
  • /models: Trained models persisted in .joblib.

🛠️ How to Run the Pipeline

  1. Clone the repository.
  2. Create a virtual environment: python -m venv .venv.
  3. Install dependencies with pinned versions: pip install -r requirements.txt.
  4. Processed datasets are already available in /data/processed/.

Lucas Nascimento Oliveira
Data Scientist: generating intelligence and business value through data.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages