Data Quality

Missing Value Pattern Analyzer

Groups rows by which fields are missing, so a broken load shows up as one pattern instead of several sparse columns.

Loading the tool…

Processing happens locally in your browser. What you paste or load is processed by this page and is not uploaded to a server. Nothing is stored unless you use a control that says it stores something, and you can clear anything this site has kept from the privacy page.

How to use this tool

  1. Paste a file that has gaps you do not understand yet.
  2. Select Find patterns.
  3. Look at the top row of the table — the largest group tells you whether the gaps are one problem or many.
  4. If one pattern covers a large share of the rows, go and look at the step that produced those rows rather than imputing anything.

What missing value analyzer does

Per-column percentages cannot answer the question you actually need answered before you decide what to do about missing data: is it scattered, or is it a block of rows where a whole section failed to arrive? Two columns each 20% empty are one problem if the same rows are empty in both and two problems if they are not, and no amount of column-level reporting distinguishes them.

This groups every row by the exact set of fields it is missing and ranks the groups by how many rows fall into each. A pattern covering a large share of the file almost always means one join or one upstream step failed rather than that the fields are independently sparse, and that is a fix at the source rather than an imputation rule.

Frequently asked questions

Because the percentages cannot tell you whether the gaps are related. Two columns each 20% empty are one broken join if the same rows are empty in both and two sparse fields if they are not, and those need entirely different responses. Grouping rows by exactly what they are missing is the only way to see which you have.

That a step failed for a subset of rows: a lookup that matched nothing, a source system that was down for part of the extract window, or a file that was concatenated from two exports with different columns. It is almost always a fix at the source rather than a case for imputation.

Yes, as its own group labelled "nothing missing". Seeing how large that group is relative to the others is the fastest way to judge whether the file is broadly sound with a few gaps or broadly incomplete.

The forty largest. Beyond that the tail is individual rows with their own combinations, which is not a pattern and not something you can act on as a group.