Chart Results and Distribution Analysis
Calculated Data
| Dataset | Category | Value | Percentage | Rank | Cumulative | Difference from Average | Status | Color |
|---|
Dataset Comparison
| Category | Dataset A | Dataset B | Difference | Percentage Change |
|---|
Copyable Result
Formula Used
Category percentage: Percentage = (Category Value ÷ Total Value) × 100.
Imbalance ratio: Largest Category Value ÷ Smallest Positive Category Value.
Distribution similarity: 100 × (1 − Total Variation Distance).
How to Use
- Choose counts or percentage input mode.
- Enter category labels and numeric values.
- Select chart, label, color, and ordering options.
- Enable comparison mode when two datasets are needed.
- Create the chart and review imbalance findings.
- Export the visual, table, or complete report.
Example Data
| ML Example | Categories | Values | Recommended View |
|---|---|---|---|
| Binary classes | Positive, Negative | 420, 180 | Donut |
| Dataset split | Train, Validation, Test | 70, 15, 15 | Standard pie |
| Customer clusters | Cluster 1, 2, 3, 4 | 120, 95, 60, 25 | Exploded pie |
| Sentiment | Positive, Neutral, Negative | 52, 28, 20 | Donut |
Frequently Asked Questions
When should a pie chart be used?
Use it for a few categories forming one meaningful total.
How many slices are recommended?
Six or fewer slices usually provide the clearest comparison.
How does this calculator detect imbalance?
It compares majority and minority values and their percentages.
Can percentage inputs be normalized?
Yes. Values can be rescaled automatically to total one hundred.
What happens to very small slices?
They can be combined automatically into an Other category.
Can actual and predicted distributions be compared?
Yes. Comparison mode provides charts, differences, and similarity.
Which export formats are supported?
The calculator supports PNG, SVG, CSV, PDF, copy, and print.
Are negative category values allowed?
No. Pie slices require zero or positive numeric values.
When is a bar chart better?
Use bars for many categories or precise ranked comparisons.