Datana
Back to home

Documentation

Quick guide to using Datana for automated data analysis.

Getting started

  • Create an account at /register
  • Upload a CSV, Excel, JSON, TSV, or Parquet file from the dashboard or import from a link or database (see below for more details)
  • Start an analysis and wait for the pipeline to complete
  • Review insights, charts, forecasts, and downloadable reports

Bringing in data

  • File upload: up to 100MB; multi-sheet Excel workbooks are merged when sheets share columns, otherwise the largest sheet is used
  • From a link: public CSV, Excel, JSON, or Parquet URLs, and Google Sheets shared with “anyone with the link”
  • From a database (paid plans): add a PostgreSQL or MySQL connection under Connections, then pick a table or write a single SELECT query and set a row limit. Use a read-only database user

Analysis options

When you start or re-run an analysis you can choose the main measure, the date column, and the model target. Leave any of them blank and Datana will detect them automatically. Results show which choices were yours and which were auto-detected.

Reading your results

  • Overview: executive summary, key findings, KPIs, risks, opportunities, and recommendations
  • Charts: filter by date range and one or more segment values; download single charts or all as a zip
  • Data quality and statistics: quality scores, issues, descriptive statistics, correlation, and covariance
  • Tests: automatic significance tests with p-values, effect sizes, and plain-language conclusions
  • Forecast: best of Holt-Winters, ARIMA, and Prophet with a 95% range (needs a date column and enough history)
  • Model: the best predictive model, how well it fits, what drives the target, and a what-if form
  • Customers: RFM segments and lifetime value when the data has customer, date, and amount columns
  • Reports: PDF, HTML, and PowerPoint, plus the cleaned dataset as CSV or Parquet

AI chat

Open chat from a completed analysis and ask in plain English. Chat can run read-only SQL on your cleaned data, aggregate and filter, compare the latest period with the one before, forecast any numeric column, and draw charts. Use “Show work” to see each step and query, and pin useful charts to the analysis.

Sharing & collaboration

  • Share → create a read-only link with an expiry (or none) and optionally allow report downloads; revoke it at any time
  • Invite teammates by email as viewers or commenters; they accept under Shared after signing in with that email
  • Comments can point at a specific tab or chart; you're emailed when someone else comments

Scheduled refresh & alerts

  • Datasets imported from a link or database show a Data source card with Refresh now and a schedule
  • Refresh daily or weekly at a chosen hour (UTC). Each run re-imports the data, re-runs the analysis, and compares it with the previous run
  • Get an email when a chosen metric drops by a set percentage, when the latest period looks unusual, or when a refresh fails
  • Starter includes 1 connection and weekly refresh; Professional and above include more connections and daily refresh

Dashboard

  • Datasets: upload, import, and view schema details and data sources
  • Analysis: track job progress and open results
  • Chat: ask questions about a completed analysis
  • Connections: manage database connections
  • Shared: analyses other people have shared with you
  • Billing: manage subscriptions and credit packs
  • Settings: update profile and password

API

Datana's web and mobile apps run on a REST API. Public API keys and developer documentation are not available yet if you need programmatic access, email [email protected].