ANALYSIS AND IMPROVEMENT

Let your data guide your business.

Your orders, customer transactions and operational records are already in your project. Describe your question and use your own project data to examine processes, measure results and prepare forecasts.

Use the data in your project

There is no need to prepare CSV files or upload your records again. Your data already lives in the connected project database; the workspace reads only the sources and fields you are authorized to use.

An analysis workspace must be connected to your project. You may need to sign in with your project account. If no connection is available, the system tells you; it never switches to another project’s data.

Start with a question

“At which step do orders wait longest?”

“Which campaigns bring better results?”

“Forecast sales for the coming weeks.”

The system reads an authorized database snapshot and runs data quality, process mining, marketing measurement, forecasting and classification flows. You review the draft and fields, then connect the result to an approved automation or report workflow.

Data qualityProcess miningForecastingMachine learningScheduled automationSQL Server · PostgreSQL · MySQL
Optional: file analysis tools

This separate tool is for external files. Use project analysis above to work with records already in your system.

What do you want to learn?

Your request helps choose an analysis type. This screen runs defined analyses, not arbitrary Python/SQL or LLM commands.

Prepare your data

Up to 1,000 records / 20 columns. Include only necessary columns; remove names, emails, passwords, keys and unnecessary personal data. All CSV columns are sent; mapping does not remove unused columns. Choosing a file does not send it.

Findings and next step

Measurements, data warnings and suggested reviews will appear here after analysis. No analysis has run yet.

What changed after the change?

Keep a measurement and its analysis recipe. Measure the next dataset with the same recipe and see the results side by side. A numeric difference alone does not prove improvement or causation.

Which data does this measurement describe?

Run an analysis above first. The scope and dates below are your declarations; they do not filter the CSV or verify its source. Do not enter secrets or personal names.

For example, Europe/Istanbul or UTC. Use the data’s time zone; dates are not automatically converted.

The file contains column names, mapping, measurements, warnings and the declared scope above. It excludes raw rows, your chat request and project credentials. You keep the file; it is not automatically saved to your account.

Choose a baseline measurement

Up to 64 KiB JSON. Choose the measurement file from this section, not an analysis report or improvement request. Opening it does not upload it.

Prepare the new CSV above, apply the recipe, then give fresh upload consent and analyze it. Column names and order must match. Opening a file does not change mapping or upload consent.

No baseline selected.