Best Data Viz
Relationship

Correlogram

Symmetric grid of every variable pair's correlation coefficient, with circle size encoding magnitude and color encoding sign — diagonal blocks of strongly-correlated variables jump out instantly.

Subject Score Correlations

30 students across 4 subjects (scores 35-99)

View data (30 rows)
Chart data table: Subject Score Correlations
MathScienceEnglishArt
61716873
84757668
54577775
68777672
69716065
76675960
77727868
67546680
84806656
86877974
81805759
79807979
75775953
64727677
85856963
83836562
62617689
69705650
76805952
85867068
72717369
77798181
65657375
70656358
72708586
68776974
76796466
40367991
92927271
71678478
Make a correlogram with your data

Use a correlogram when…

  • Exploring relationships among many variables
  • Feature selection in ML

Avoid when…

  • Non-numeric data
  • Fewer than 3 variables

Data it needs

PropertyValue
Min Rows10
Min Columns3
Column Types
numbernumbernumber

Visual anatomy

Marks
circletext
Channels
color-divergingsizeposition
Axes
variable matrix

Guiding principles

Consider instead

Common mistakes

  • Assuming correlation is causation

  • Including too many variables

History

Standard technique in multivariate statistics since Pearson's correlation work.

Accessibility notes

Provide correlation coefficient values as text. Use diverging color with labels.

Related reading

Got data? Let's see what works.

Drop your CSV. You'll get a Correlogram plus four alternatives - ranked by which one actually fits your data best.