Enter Dataset
Input preview
This browser preview shows up to six rows before analysis.
Formula Used
Missing percentage = Missing values ÷ Total values × 100.
Column completeness = Available values ÷ Column rows × 100.
Row completeness = Available cells in row ÷ Total columns × 100.
Dataset retention = Remaining cells ÷ Original cells × 100.
How to Use
- Paste tabular data or upload a supported file.
- Choose the delimiter and define missing-value markers.
- Submit the form to generate the interactive analysis.
- Filter rows and columns using heatmap controls.
- Preview cleaning strategies without changing the source data.
- Export reports, processed data, or missing-cell coordinates.
Example Data
| Customer_ID | Age | Income | Region | Join_Date | Satisfaction | Churn |
|---|---|---|---|---|---|---|
| C001 | 34 | 52000 | North | 2025-01-14 | 8 | No |
| C002 | 29 | South | 2025-02-03 | 7 | No | |
| C003 | 41 | 68000 | East | 9 | No |
Understanding Missing Data
MCAR
Missing Completely at Random means missingness has no systematic cause. Simple methods may remain reasonable under this assumption.
MAR
Missing at Random depends on other observed variables. Conditional modelling can often reduce resulting bias.
MNAR
Missing Not at Random depends on unobserved values themselves. Domain investigation and sensitivity analysis become especially important.
Frequently Asked Questions
What does a missing-value heatmap show?
It maps available and missing cells across rows and columns. Patterns become easier to detect visually.
Which values count as missing?
Blank cells, NULL, NA, N/A, NaN, and custom markers count. Matching is case-insensitive.
Should every incomplete row be deleted?
No. Deletion can reduce sample size and introduce bias. Review the missingness mechanism first.
When is mean imputation suitable?
Mean imputation can work for roughly symmetric numerical variables. It reduces variance and may distort relationships.
Why add missing indicators?
Indicators preserve whether a value was originally missing. Some models can learn useful missingness signals.
Can categorical columns use numerical imputation?
Usually not. Mode or explicit missing-category replacement is generally clearer for categorical data.
What does co-missingness mean?
It measures how often two columns are missing together. Strong overlap may reveal collection issues.
Does this calculator modify uploaded files?
No. Analysis and treatment previews happen in memory. Your original file remains unchanged.
Can large datasets be analyzed?
Yes, within browser limits. Sampling controls keep large heatmaps responsive and readable.