ADA: Automated Data Analyst
ADA: Automated Data Analyst turns CSV and Excel files into dashboards, executive briefs, anomaly flags, forecasts, and answers to questions written in plain …
What ADA: Automated Data Analyst is
ADA: Automated Data Analyst turns CSV and Excel files into dashboards, executive briefs, anomaly flags, forecasts, and answers to questions written in plain English. Its dashboard presents trends, segment breakdowns, movement waterfalls, and segment-by-period heatmaps. Ask ADA answers questions using pandas calculations and displays the calculation with the answer. It flags periods outside a calibrated band and uses a guarded forecasting baseline that can decline to run when history is thin or report when it does no better than assuming no change. Users can download a Markdown executive brief and a cleaned CSV. ADA accepts .csv, .xlsx, and .xlsm files up to 25 MB and 250,000 analyzed rows. It can run locally as a Streamlit app from its source repository without an API key; local processing includes file parsing, cleaning, charts, forecasting, and supported questions. Optional AI calls require a key and receive schema and computed evidence rather than uploaded rows, though evidence may include segment labels. ADA is MIT licensed and may be used commercially with the copyright notice retained.
Good
- Shows pandas calculations behind plain-English answers
- Exports Markdown briefs and cleaned CSV files
- Runs locally without an API key
- Accepts CSV, XLSX, and XLSM files
Watch out
- Files are limited to 25 MB
- Analysis is limited to 250,000 rows
- Optional AI calls require a key
- Hosted app is not approved for regulated data
Who it suits
ADA suits people who want to explore CSV or Excel data, inspect calculations, and export a brief or cleaned file. Self-hosting is an option for users who want listed analysis tasks computed locally.
Verdict
ADA combines dashboarding, explanations, anomaly flags, forecasting, and exports for spreadsheet data. Users handling regulated data should note the hosted-app caveat and consider the listed local option.
The facts
Each one read from ADA: Automated Data Analyst's own pages, with the date it was checked.
What it does
- AI safeguards
- ADA validates model output against a typed schema and the actual data, and it never executes model-generated code.checked 7 Oct 2026
- Anomaly detection
- Anomaly flags identify periods outside a calibrated band, sized so a stable series false-alarms about once in twenty analyses.checked 7 Oct 2026
- Ask ADA
- Plain-English questions are answered with pandas calculations, and the calculation is shown with the answer.checked 7 Oct 2026
- Calculations
- ADA shows the calculation behind its findings and labels recommendations as interpretation rather than cause.checked 7 Oct 2026
- Dashboard
- The dashboard includes trends, segment breakdowns, a movement waterfall, and a segment-by-period heatmap.checked 7 Oct 2026
- Downloads
- ADA can download a Markdown executive brief and a cleaned CSV.checked 7 Oct 2026
- Exports
- Users can download a Markdown executive brief and a cleaned CSV.checked 7 Oct 2026
- File limits
- The stated limits are 25 MB per file and 250,000 analyzed rows; supported formats are .csv, .xlsx, and .xlsm.checked 7 Oct 2026
- Forecasting
- Forecasting uses a guarded baseline that declines to run on thin history and reports when it does no better than assuming no change.checked 7 Oct 2026
- Hosted demo privacy
- The hosted demo holds uploaded files in session memory and does not write them to a database.checked 7 Oct 2026
- Integration
- The maker describes an optional OpenAI integration for planning questions the rule parser cannot handle; pandas executes the resulting calculation locally.checked 7 Oct 2026
- License
- ADA is MIT licensed, and the README says it may be used, changed, and shipped commercially with the copyright notice retained.checked 7 Oct 2026
- Local processing
- When self-hosted, file parsing, cleaning, charts, forecasting, and supported question answering run locally with pandas and no network call.checked 7 Oct 2026
- Local use
- ADA can be run as a Streamlit app from its source repository, and the README says no API key is required.checked 7 Oct 2026
- Optional AI
- Optional, button-triggered AI calls require a key and receive schema and computed evidence rather than uploaded rows; the README notes evidence may include segment labels.checked 7 Oct 2026
- Privacy
- When run locally without an API key, file parsing, cleaning, schema detection, charts, anomaly detection, forecasting, and Ask ADA are computed locally.checked 7 Oct 2026
- Purpose
- ADA analyzes CSV and Excel files to create a dashboard, executive brief, anomaly flags, forecast, and answers to plain-English questions with calculations shown.checked 7 Oct 2026
- Security
- The security policy says hosted infrastructure processes network traffic and application memory, and cautions that the public hosted app is not thereby approved for regulated data.checked 7 Oct 2026
- Security limitation
- The maker says the hosted app is not thereby an approved environment for regulated data.checked 7 Oct 2026
- Support
- The README directs users to documentation, contributing guidance, GitHub issues, and the project roadmap.checked 7 Oct 2026
The company
- Maker
- The repository credits Sainesh Nakra as the maker, and his site lists San Jose, California as his location.checked 7 Oct 2026
