ADA is an open-source automated data analyst from GitHub user saineshnakra that turns CSV and Excel files into interactive Plotly dashboards with anomaly detection, guarded forecasts, and plain-English Q&A. Upload a spreadsheet and ADA cleans it, infers the business schema, calculates trends and drivers, and surfaces prioritized investigations — all with deterministic, traceable math shown for every answer.

Unlike typical chat-with-CSV tools, ADA keeps four layers explicit: deterministic calculation for trends and anomalies, auditable conversation that executes pandas query plans locally, labeled interpretation for recommended actions, and optional AI that only sees schema and computed evidence. No uploaded rows ever reach a model. The engine runs fully without an API key; when configured, it uses OpenAI's gpt-5.6-luna for strategic reads and gpt-5.6-terra for ambiguous queries, both typed and cached.

Built with Streamlit and pandas, ADA is MIT-licensed and free. The architecture separates pure analysis functions from UI and model boundaries, making the business engine testable without network access. Run locally with `streamlit run app.py` — no secrets required unless enabling optional AI features.