💡 What is a 2D density chart?

There are several chart types allowing to visualize the distribution of a combination of 2 numeric variables. They always have a variable represented on the X axis, the other on the Y axis, like for a scatterplot (left).

Then the number of observations within a particular area of the 2D space is counted and represented with a color gradient. The shape can vary: hexagones result in a hexbin chart, squares in a 2d histogram. A kernel density estimate can be used to get a 2d density plots or a contour plots

Cheat sheet: line customization with matplotlib

Confusing? Visit data-to-viz to clarify..

Contour plot with Seaborn

The contour plot can be easily built thanks to the kdeplot() function of the Seaborn library.

Basic contourplot with seaborn.

Basic contourplot with seaborn.

Basic contourplot with seaborn.

2D histogram with Matplotlib

Build a 2d histogram thanks to the hist2d() function of the Matplotlib library. Do not forget to play with the bins argument to find the value representing the best your data.

Basic 2D histogram with bins customization

Control the color in the 2d histogram.

Black and white, color scale legend.

Hexbin chart with Matplotlib

Split the graph area in hexagones and you get a hexbin density chart. This time, it is matplotlib that gets you covered thanks to its hexbin() function.

Basic hexbin with bins customization

Control the color in the hexbin.

Black and white, color scale legend.

2d density chart with Matplotlib

2D densities are computed thanks to the gaussian_kde() function and plotted thanks with the pcolormesh() function of matplotlib().

Basic 2d density with bins customization

Control the color in the 2d density.

Black and white, color scale legend.

2d density and marginal plots

2D densities often combined with marginal distributions. It helps to highlight the distribution of both variables individually. It is pretty straightforward to add thanks to the jointplot() function of the Seaborn library.

Marginal distribution and hexbin.

Marginal distribution and contour.

General customization.