Complete your first data task

Open a table, check data quality, generate a plan, and execute and review the result.

For a first pass, choose one small task: identify missing values, ask what a table contains, or create a basic cleaning plan for one table.

The goal is to learn the loop: open data, inspect it, ask a focused question, review the answer or plan, then decide whether to clean, export, or continue exploring.

Open table

Select a table from the project sidebar or table tabs. The central grid shows rows and columns, with headers that expose field-level controls.

If the table is large, IMD may show a practical preview rather than every row at once. Use the table view to understand shape before asking AI for interpretation.

Check data quality

Look for empty cells, inconsistent categories, suspicious values, and ranges that do not match your expectations. Quality indicators and diagnostics panels help surface these issues.

Do not fix everything immediately. Note which fields are important for your analysis and focus first on issues that change the meaning of results.

Generate plan

Use the Cleaning tab to prepare an editable cleaning plan for the current table. The plan should describe proposed steps before data is modified.

A good plan is specific: it names fields, describes the issue, and explains what the output should look like after the step runs.

Execute and review

Run the plan only after reviewing the steps. After a run, compare the cleaned result with the original table. Check row count, field names, important categories, missing values, and any field used in later analysis.

Do not treat a successful run as automatic approval. The user remains responsible for validating whether the output is analytically correct.