Advanced Analytic Tool

A lightweight, notebook-friendly analytical framework built to speed up exploratory data analysis and rapid prototyping of models for structured datasets. The tool emphasizes:

  • Unified data interface for tabular, time–series, and semi‑structured sources.
  • Pluggable pipeline components (cleaning, imputation, encoding, feature selection).
  • Automatic experiment tracking (metrics + parameters) with reproducible config files.
  • Model zoo covering tree ensembles, linear models, and compact neural architectures.
  • Statistical and SHAP-based interpretability reports.

Tech: Python, Pandas, Scikit‑learn, LightGBM, FastAPI (for optional service layer), Docker.

Status: Internal research utility evolving toward an open-source developer toolkit.

Includes reproducible config files and deterministic seeding to ensure comparable experiment runs across environments.

View on GitHub →