01 / FACETED VIOLINS
Compare the shape of each group Four variants, four stages and 48 observations per stage. Common axes make the small multiples comparable.
Variant A
Variant A: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
Stage 1
Stage 2
Stage 3
Stage 4
0
20
40
60
Example units
Variant A: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
1
2
3
4
Stage
0
10
20
30
40
50
60
Example units
Variant B
Variant B: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
Stage 1
Stage 2
Stage 3
Stage 4
0
20
40
60
Example units
Variant B: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
1
2
3
4
Stage
0
10
20
30
40
50
60
Example units
Variant C
Variant C: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
Stage 1
Stage 2
Stage 3
Stage 4
0
20
40
60
Example units
Variant C: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
1
2
3
4
Stage
0
10
20
30
40
50
60
Example units
Variant D
Variant D: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
Stage 1
Stage 2
Stage 3
Stage 4
0
20
40
60
Example units
Variant D: four synthetic distributions Forty-eight generated observations per stage. Each violin shows a kernel density estimate, clipped to the observed range, with quartile lines. All facets use the same zero-to-sixty vertical scale; width is normalised within each violin.
1
2
3
4
Stage
0
10
20
30
40
50
60
Example units
Density describes shape; inner lines show quartiles. Violin widths are normalised per group and do not encode sample size.
02 / ANNOTATED HEATMAP
Find structure in a matrix Six invented rows across six illustrative periods. Printed values preserve the detail behind the colour.
Annotated synthetic counts across six rows and six periods Thirty-six independently generated integer counts. Every cell prints its value; darker green means a higher count. Labels and relationships do not correspond to actual workplace records.
P1
P2
P3
P4
P5
P6
Illustrative period
Row A
Row B
Row C
Row D
Row E
Row F
3
3
6
3
3
5
6
8
7
25
10
9
7
7
9
4
4
8
13
15
19
13
17
24
5
22
3
5
1
10
9
12
10
11
5
15
0
5
10
15
20
25
Synthetic count
Annotated synthetic counts across six rows and six periods Thirty-six independently generated integer counts. Every cell prints its value; darker green means a higher count. Labels and relationships do not correspond to actual workplace records.
P1
P2
P3
P4
P5
P6
Illustrative period
A
B
C
D
E
F
3
3
6
3
3
5
6
8
7
25
10
9
7
7
9
4
4
8
13
15
19
13
17
24
5
22
3
5
1
10
9
12
10
11
5
15
0
5
10
15
20
25
Synthetic count
All 36 cell values are synthetic. The matrix contains 336 counts in total; colour intensity uses a common scale.
03 / HISTOGRAM · KDE · ECDF
Read one distribution three ways The same 300 generated observations appear in all three views.
Histogram
Histogram of the same 300 synthetic observations The histogram, kernel-density curve and empirical cumulative distribution use exactly the same generated observations. The density curve uses Scott bandwidth; the cumulative curve counts observations at or below each value.
0
10
20
30
40
Example units
0.00
0.02
0.04
Density
Histogram of the same 300 synthetic observations The histogram, kernel-density curve and empirical cumulative distribution use exactly the same generated observations. The density curve uses Scott bandwidth; the cumulative curve counts observations at or below each value.
0
10
20
30
40
Example units
0.00
0.02
0.04
Density
Kernel density
Kernel density of the same 300 synthetic observations The histogram, kernel-density curve and empirical cumulative distribution use exactly the same generated observations. The density curve uses Scott bandwidth; the cumulative curve counts observations at or below each value.
0
10
20
30
40
Example units
0.00
0.02
0.04
Density
Kernel density of the same 300 synthetic observations The histogram, kernel-density curve and empirical cumulative distribution use exactly the same generated observations. The density curve uses Scott bandwidth; the cumulative curve counts observations at or below each value.
0
10
20
30
40
Example units
0.00
0.02
0.04
Density
Empirical cumulative distribution
Empirical cumulative distribution of the same 300 synthetic observations The histogram, kernel-density curve and empirical cumulative distribution use exactly the same generated observations. The density curve uses Scott bandwidth; the cumulative curve counts observations at or below each value.
0
10
20
30
40
Example units
0%
50%
100%
Share ≤ x
Empirical cumulative distribution of the same 300 synthetic observations The histogram, kernel-density curve and empirical cumulative distribution use exactly the same generated observations. The density curve uses Scott bandwidth; the cumulative curve counts observations at or below each value.
0
10
20
30
40
Example units
0%
50%
100%
Share ≤ x
Bins show frequency density; KDE smooths with Scott bandwidth; ECDF gives the observed share at or below a value.
04 / PERCENTILE COMPARISON
Locate the upper tail Two new samples of 100 observations, shown on common axes. Dashed guides locate each 90th percentile.
Variant A
Variant A: linearly interpolated percentiles One hundred independently generated observations, sorted by value. Plotting positions run from zero to one hundred in steps of one hundred divided by n minus one, matching linear quantile interpolation. Guides meet the curve at the 90th percentile. Both variants use common axes, from zero to ninety example units.
0
20
40
60
80
Example units
0%
50%
90%
100%
Interpolated percentile
P90 = 36.2
Variant A: linearly interpolated percentiles One hundred independently generated observations, sorted by value. Plotting positions run from zero to one hundred in steps of one hundred divided by n minus one, matching linear quantile interpolation. Guides meet the curve at the 90th percentile. Both variants use common axes, from zero to ninety example units.
0
20
40
60
80
Example units
0%
50%
90%
100%
Interpolated percentile
P90 = 36.2
Median 24.5 · P90 36.2 Variant B
Variant B: linearly interpolated percentiles One hundred independently generated observations, sorted by value. Plotting positions run from zero to one hundred in steps of one hundred divided by n minus one, matching linear quantile interpolation. Guides meet the curve at the 90th percentile. Both variants use common axes, from zero to ninety example units.
0
20
40
60
80
Example units
0%
50%
90%
100%
Interpolated percentile
P90 = 49.1
Variant B: linearly interpolated percentiles One hundred independently generated observations, sorted by value. Plotting positions run from zero to one hundred in steps of one hundred divided by n minus one, matching linear quantile interpolation. Guides meet the curve at the 90th percentile. Both variants use common axes, from zero to ninety example units.
0
20
40
60
80
Example units
0%
50%
90%
100%
Interpolated percentile
P90 = 49.1
Median 32.0 · P90 49.1 Plotting positions and quantiles use the same linear interpolation convention. These are illustrative distributions, not workforce benchmarks.
Four examples, complementary questions. All observations, matrix values and plotted relationships were generated independently for this gallery. Matplotlib and Seaborn render the figures; NumPy and SciPy calculate the recorded statistics. The original gallery informs chart selection and presentation only.