Pinned Loading

1. analytics-pipeline analytics-pipeline Public

   Reusable data pipeline: load, clean, validate, store (DuckDB), analyze (SQL), report. Built with Python and pandas, with tests.

   Python

2. claim-denial-analysis claim-denial-analysis Public

   Which claims get denied, what denials cost, and what a billing team should do: DuckDB SQL + pandas + Streamlit

   Python

3. ml-pipeline-freMTPL2-claims ml-pipeline-freMTPL2-claims Public

   16-step ML pipeline: car-insurance claim frequency on 678K real policies (Poisson GLM, gradient boosting, XGBoost)

   Python

4. ml-pipeline-pricing-promo ml-pipeline-pricing-promo Public

   Predicting which discounted order lines lose money: a 16-step ML pipeline (scikit-learn, XGBoost) with leakage checks and tests.

   Python

5. pricing-promo-analysis pricing-promo-analysis Public

   Which discounts raise revenue and profit, and which just give away margin: DuckDB SQL + pandas + statsmodels + Streamlit

   Python

6. review-insights-ai review-insights-ai Public

   What customers complain about and what each complaint costs in stars: local llama3.2 (Ollama) labelling + DuckDB SQL + statsmodels + Streamlit

   Python