Dear all,
Is there a way to adapt the "fontsize" parameter for labels as a function of boxes sizes?

Hi, I was working on a project where I wanted to use squarify to visualize my dataset. I was wondering if subgrouping is available.

For subgrouping I mean something like this:
image

I can see that we can achieve the first level grouping(GROCERY I, DELI, CLEANING, ...), but it seems that we can't further group items within each category.

Please let me know if I'm understanding this correctly, thanks

I am running the code below from the gist "https://gist.github.com/gVallverdu/0b446d0061a785c808dbe79262a37eea":

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import pandas as pd
import squarify

df = pd.read_csv("C:/PProjects/structure_population.csv", sep=";")
df = df.set_index("libgeo")
df = df[["superf", "p11_pop"]]
df2 = df.sort_values(by="superf", ascending=False)

x = 0.
y = 0.
width = 100.
height = 100.
cmap = matplotlib.cm.viridis

mini, maxi = df2.drop("PAU").p11_pop.min(), df2.drop("PAU").p11_pop.max()
norm = matplotlib.colors.Normalize(vmin=mini, vmax=maxi)
colors = [cmap(norm(value)) for value in df2.p11_pop]

labels = ["%s\n%d km2\n%d hab" % (label) for label in zip(df2.index, df2.superf, df2.p11_pop)]
labels[11] = "MAZERES-\nLEZONS\n%d km2\n%d hab" % (df2["superf"]["MAZERES-LEZONS"], df2["p11_pop"]["MAZERES-LEZONS"])

fig = plt.figure(figsize=(12, 10))
fig.suptitle("Population et superficie des communes de la CAPP", fontsize=20)
ax = fig.add_subplot(111, aspect="equal")
ax = squarify.plot(df2.superf, color=colors, label=labels, ax=ax, alpha=.7)
ax.set_xticks([])
ax.set_yticks([])
ax.set_title("L aire de chaque carre est proportionnelle a la superficie de la commune\n", fontsize=10)

img = plt.imshow([df2.p11_pop], cmap=cmap)
img.set_visible(False)
fig.colorbar(img, orientation="vertical", shrink=.96)

fig.text(.76, .9, "Population", fontsize=10)
fig.text(.5, 0.1,
"Superficie totale %d km2, Population de la CAPP : %d hab" % (df2.superf.sum(), df2.p11_pop.sum()),
fontsize=14,
ha="center")
fig.text(.5, 0.07,
"Source : http://opendata.agglo-pau.fr/",
2 fontsize=14,
ha="center")
plt.savefig('foo.png')
`

If I plot manually using rects coordinates , then it render correctly. Tried on both python 2.7/3/6 , Windows, PyCharm for GUI coding, backend for matplotlib is TkAgg, changing to Agg /creating image did not make difference.

My output plot is below

Treemap is a great package to use.

Is there any way to adjust the font size inside each treemap block?

And is there any way to handle label text outside of the block?

Using the most recent version, here's a MWE:

import squarify as sq

data = [1, 2, 3]
labels = ["1", "2", "3"]
sq.plot(sizes=data, label=labels)

This throws the error

Traceback (most recent call last):
  File "/Users/nraf1041/Desktop/charting/test/test_competitive.py", line 6, in <module>
    competitive(dpis)
  File "/Users/nraf1041/Desktop/charting/charting/partisan/competitive.py", line 59, in competitive
    sq.plot(sizes=data, label=labels)
  File "/Users/nraf1041/.pyenv/versions/3.8.2/lib/python3.8/site-packages/squarify/__init__.py", line 268, in plot
    ax.text(x + dx / 2, y + dy / 2, l, va=va, ha="center", **text_kwargs)
TypeError: text() missing 1 required positional argument: 's'

Is it possible to get all boundaries to touch (without padding)? When I round to integers boxes borders are inconsistent.

A small issue, just to clarify one sentence in the example provided:
I had already made some apparently successful test-plots of unsorted data, before realizing of this sentence in the example provided:

# values must be sorted descending (and positive, obviously)
values.sort(reverse=True)

As my unsorted plot looks OK, I guess that "must" was just a statement of the intended layout for that example (not mandatory for squarify to work properly).

As for my plot layout, I am guessing too that squarify will always create boxes in the provided order, from lower left corner upwards, and them from left to right. Correct? Or may squarify sometimes change that order to achieve the best fit of boxes?

It is interesting to know if our provided order permits to control layout of consecutively generated graphs, with slight differences in values, in order to later generate a nice-looking animated graphs.
This example (historic population comparisons between countries ) illustrates what I mean:

Taken from https://wilkox.org/treemapify/#animated-treemaps

Hi, some people find your library very useful.

Do you plan to maintain it? i.e. make releases a bit more often?

I found some forks, for instance, the one from @ocket8888 seems to be adding fixes:
https://github.com/Sensibility/squarify

Perhaps it makes sense for you and them to work together in one repo? and help the community along this collaboration ;)

Probably this is not a squarify issue but some matplotlib-related help requests.
As they might be interesting for other squarify-matplotlib new users like me, I dare to ask them here:

(1) When using plot, is it possible to draw boxes with a tiny one pixel border of a chosen contrasting colour (i.e., a black and white dashed or dotted line, so it will be always visible)?

When I run the example below (based in #26), boxes colors are random and I am OK with that ... but sometimes two neighbour colors are so similar I can't tell them apart.
(2) I know I can use padding, but then coloured boxes are not of their real sizes (that is specially relevant for the smallest sized boxes: see E box below, which represents 5 square units but looks like 1-sized in the padded graph). A tiny line showing the real box limits would improve visibility without producing this wrong box size impression.

	import matplotlib.pyplot as plt
	import squarify
	squarify.plot(
		label=["A", "B", "C", "D","E","F","G","H"], # list-like used as label text
		sizes=[25, 13, 22, 35, 5, 250, 650, 24], # sum = 1024
		norm_x=32, norm_y=32,  # product = 1024
		pad=False, # white gap from each rectangle border towards its center
		)
	plt.show()

In my plot, A-B-C-D-E are all drawn in a left column, whereas F-G-H occupy most of the remaining area.
If you run the script several times, there are chances that random colors are too close to each other, so square boundaries in that left column are almost invisible (see B-C-D or F-H boundaries below).

(3) Another question which you can also see in these examples.
I would expect a 32x32 square graphic, but I see rectangles. Looks like axes ratios are being modified somehow. How can I control this?

Thanks a lot for this great package !!
@abubelinha

Is there a way to change color and size of the font?

Hello,
With squarify, I try to do an treemap but I have an :

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-142-f809b7b5caff> in <module>
     40 
     41 # Plot
---> 42 squarify.plot(sizes='data' , label='labels' , alpha=.7)
     43 plt.axis('off')
     44 plt.show()

/opt/jupyterhub/lib/python3.7/site-packages/squarify/__init__.py in plot(sizes, norm_x, norm_y, color, label, value, ax, pad, bar_kwargs, text_kwargs, **kwargs)
    239         bar_kwargs.update(kwargs)
    240 
--> 241     normed = normalize_sizes(sizes, norm_x, norm_y)
    242 
    243     if pad:

/opt/jupyterhub/lib/python3.7/site-packages/squarify/__init__.py in normalize_sizes(sizes, dx, dy)
    168         The normalized values.
    169     """
--> 170     total_size = sum(sizes)
    171     total_area = dx * dy
    172     sizes = map(float, sizes)

TypeError: unsupported operand type(s) for +: 'int' and 'str'

My code is :

from SPARQLWrapper import SPARQLWrapper, JSON
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import re

sparql = SPARQLWrapper("http://isidore.science/sparql")
sparql.setQuery("""
PREFIX sioc: <http://rdfs.org/sioc/ns#>
PREFIX dcterms: <http://purl.org/dc/terms/>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
PREFIX ore: <http://www.openarchives.org/ore/terms/>
PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT ?label (count(?documents) as ?count) WHERE {
    <http://isidore.science/collection/10670/2.qfy8eq> ore:aggregates ?documents.
    ?documents sioc:topic ?topic.
    ?topic skos:prefLabel ?label.
    FILTER(lang(?label)="fr")
}
GROUP BY ?label
ORDER BY DESC(?count) LIMIT 10

""")
sparql.setReturnFormat(JSON)
results = sparql.query().convert()
results_df = pd.json_normalize(results['results']['bindings'])
labels = results_df['label.value']
data = results_df['count.value']

# Dataframe
df = pd.DataFrame({'data':data, 'labels':labels})
print(df)

# Tree
squarify.plot(sizes='data' , label='labels' , alpha=.7)
plt.axis('off')
plt.show()

What's wrong with this data in a DataFrame ?
Best,
Stéphane.

It will be useful if the user could pass a colormap of his/her choice when plotting.

It seems like it could be done with a simple change around line 134. I'll be happy to make a PR. Is that okay?

Release 0.4.0 breaks passing matplotlib.plot() kwargs in the squarify.plot() method.
This worked in v0.3.0.
This feature is mainly useful for plotting squarify output on matplotlib figures.
For instance, in the code example below I use a figure with fixed size to plot a treemap on.

import squarify

def plot_type_treemap_matplot(type_df, fp="type_treemap_matplot.png"):

    type_df["label"] = type_df.apply(lambda x: f"{x.name}\n{x['pct']}% (n={x['n'].astype('int')})", axis=1)

    figsize = [9, 5]
    plt.rcParams["figure.figsize"] = figsize
    cmap = plt.get_cmap("tab20", lut=len(type_df.index))
    # Change color
    fig = plt.figure(figsize=figsize, dpi=300)

    squarify.plot(sizes=type_df["n"], label=type_df["label"],
                  color=cmap.colors, alpha=.4, figure=fig)  # This does not work in v0.4
    plt.title("Distribution of event categories in SENTiVENT English corpus.", fontsize=12, figure=fig)
    plt.axis('off', figure=fig)
    plt.show()
    fig.savefig(fp)

In v0.4.0 this introduced AttributeErrors for the "figure" and "alpha" kwargs, while it produces the expected output in v0.3.0.

I'm building a simple treemap using the library. Is there a way to automatically resize/wrap the label of the single square?
I'm using the following line of code:
squarify.plot(label=labels,sizes=dataGoals.Streams, color = colors, alpha=.7, bar_kwargs=dict(linewidth=0.5, edgecolor="#222222"),text_kwargs={'fontsize':20})
The fontsize unfortunately is fixed, and my labels will overlap between the smaller squares:

Is there a way to resize the labels?

Thanks for the lib, while reading the code on your repo, I noticed the "label" argument can be found twice in the Docstring of the .plot method. Here's the second occurence (the 1st is a few lines before):

is it possible to change the font size of label in every box based on its value?

Thank for great library. I was able to install the library using pip, but pip is the default environment, I would like to install it to some environment eg. "myenv" in conda using:
conda install -n myenv -c conda-forge squarify

But, this commands fails.
How to install this package for conda environment?

Hello! I translated your algorithm in Java, so that it can be used as a Processing library. Source is over there https://github.com/agatheblues/squarify and library documentation is over there https://agatheblues.github.io/squarify/.

I mention you in the docs as original author, but before I propose this as a new library for Processing, I want to know if you agree with the idea, and if you want to be mentionned elsewhere?

Thanks!
Agathe

I used the pad =True. But the gap is biger. So, I want to know this parameter kan be changed to a float or int value?

Any idea why I receive the error AttributeError: module 'squarify' has no attribute 'plot'?

import matplotlib.pyplot as plt
import squarify

squarify.plot(
    sizes=[13,22,35,5],
    label=["group A", "group B", "group C", "group D"],
    )
 
plt.axis('off')
plt.show()