Distribution
ECDF
Empirical cumulative distribution function, a step function showing the proportion of data at or below each value.
API Response Time ECDF
Cumulative proportion of 50 requests
View data (50 rows)
| Response Time |
|---|
| 38 |
| 42 |
| 45 |
| 48 |
| 51 |
| 54 |
| 57 |
| 60 |
| 62 |
| 65 |
| 67 |
| 70 |
| 72 |
| 75 |
| 78 |
| 81 |
| 84 |
| 87 |
| 90 |
| 93 |
| 96 |
| 100 |
| 104 |
| 108 |
| 112 |
| 116 |
| 120 |
| 125 |
| 130 |
| 136 |
| 142 |
| 148 |
| 155 |
| 163 |
| 172 |
| 182 |
| 194 |
| 208 |
| 224 |
| 242 |
| 263 |
| 288 |
| 318 |
| 354 |
| 398 |
| 452 |
| 520 |
| 610 |
| 730 |
| 920 |
Use an ecdf when…
- Precise distribution comparison
- Finding percentiles
- Statistical analysis
Avoid when…
- General audiences unfamiliar with CDFs
- When shape/density is the focus
Data it needs
| Property | Value |
|---|---|
| Min Rows | 10 |
| Min Columns | 1 |
| Column Types | number |
Visual anatomy
Marks
step-line
Channels
position-xposition-y
Axes
x-quantitativey-proportion (0-1)
Guiding principles
Common mistakes
Not labeling the y-axis as proportion
Comparing too many groups on one plot
Pre-binning the data (an ECDF should be plotted from raw values, not a histogram)
History
Foundation of non-parametric statistics, formalized by the Glivenko-Cantelli theorem (1933) which proves the empirical CDF converges uniformly to the true distribution.
Accessibility notes
Provide summary statistics as text alternative.
Related reading
Got data? Let's see what works.
Drop your CSV. You'll get an ECDF plus four alternatives - ranked by which one actually fits your data best.