squarify | Pure Python implementation of the squarify treemap layout | Learning library

I ran into the same problem. Either of these options should work:

1. Remove any 0 values in advance, as mentioned in the comments. For a Pandas dataframe, this can be accomplished by filtering for rows that do NOT have 0 values prior to plotting. (Currently commented out in code below). OR,
2. If you want to retain the label that had a value of 0, you could add a very small fraction to all of your sizes.

You can create a list of labels by looping simultaneously through both columns and composing combined strings. For example:

Usually a single matplotlib text only has one size and one style. Here the label is created as one string with a newline.

You can use

The rectangles are stored together in a BarContainer. By default, matplotlib supposes one legend label for the complete container. To have a legend label for each individual rectangle, you can pass the BarContainer as handles to plt.legend().

The sample code below explicitly assigns colors, as the default colors can be bit hard to distinguish.

The following approach first loops through the rectangles and saves the positions of rectangles A and B.

Then, A's width and height are modified to the square root of their product.

A new L-shaped polygon is created based on the coordinates of A and B, and copying the properties of B. Afterwards, B is removed.

Note that the following code only works for the specific positions of A and B. Other positions will need a bit different L-shape, or might be impossible.

Also note that for a square to look like a square, an equal aspect ratio is needed. Squarify uses parameters norm_x and norm_y, defaulting to 100, which are the dimensions of the surrounding rectangle. You can change them to have a different aspect ratio, but be aware that this might change the positioning of the generated rectangles.

The source of the plot function looks like this:

I found the solution. Above code was almost correct expect 1 issue. Before Squerify's plot there was an statement to call show method. After Removing its working. Credit : @JahanC

Yes it is possible. You will have to write the code to extract the exact positions where you want to place the new plot in.

You need to set the position of the new figure using f.canvas.manager.window.SetPosition

This answer will greatly help https://stackoverflow.com/a/37999370/4551984

The mplcursors library can be used to create custom annotations while hovering. Here is an example with a tree map: