Download this notebook from GitHub (right-click to download).

HeatMap Element
import numpy as np
import holoviews as hv
from holoviews import opts

HeatMap visualises tabular data indexed by two key dimensions as a grid of colored values. This allows spotting correlations in multivariate data and provides a high-level overview of how the two variables are plotted.

Note: Some browsers with ad-blocking software are unable to render files named "HeatMap.ipynb". If you encounter this error locally, you can avoid it by renaming this file temporarily, or by disabling the ad blocker.

The data for a HeatMap may be supplied as 2D tabular data with one or more associated value dimensions. The first value dimension will be colormapped, but further value dimensions may be revealed using the hover tool.

data = [(i, chr(97+j),  i*j) for i in range(5) for j in range(5) if i!=j]
hm = hv.HeatMap(data).sort()

For a coarse grid like this, it can be useful to overlay the actual values onto the heatmap cells:

hm * hv.Labels(hm).opts(padding=0)

It is important to note that the data should be aggregated before plotting as the HeatMap cannot display multiple values for one coordinate and will simply use the first value it finds for each combination of x- and y-coordinates (causing a warning).

heatmap = hv.HeatMap([(0, 0, 1), (0, 0, 10), (1, 0, 2), (1, 1, 3)])
heatmap + heatmap.aggregate(function=np.max)
WARNING:param.HeatMapPlot01776: HeatMap element index is not unique,  ensure you aggregate the data before displaying it, e.g. using heatmap.aggregate(function=np.mean). Duplicate index values have been dropped.

As the above example shows before aggregating the second value for the (0, 0) is ignored unless we aggregate the data first.

To reveal the values of a HeatMap we can either enable a colorbar or add a hover tool. The hover tools even allows displaying any number of additional value dimensions, providing additional information a static plot could not capture.

Note that a HeatMap allows mixtures of categorical, numeric and datetime values along the x- and y-axes:

heatmap = hv.HeatMap((np.random.randint(0, 10, 100), np.random.choice(['A', 'B', 'C', 'D', 'E'], 100), 
                      np.random.randn(100), np.random.randn(100)), vdims=['z', 'z2']).sort().aggregate(function=np.mean)

heatmap.opts(opts.HeatMap(tools=['hover'], colorbar=True, width=325, toolbar='above', clim=(-2, 2)))

For full documentation and the available style and plot options, use hv.help(hv.HeatMap).

This web page was generated from a Jupyter notebook and not all interactivity will work on this website. Right click to download and run locally for full Python-backed interactivity.

Download this notebook from GitHub (right-click to download).