Waffle chart in time series

This post explains everything you need to know to create a waffle chart used as a time series and how to apply advanced customization. This post explains step-by-step how to replicate it with reproducible code and explanations, using pywaffle, matplotlib and other libraries.

About

This plot is a waffle chart showing the evolution of the number of animals adopted at the Long Beach animal shelter.

The chart was made by Joseph Barbier. Thanks to him for accepting sharing his work here!

Let's see what the final picture will look like:

small multiple waffle chart

Libraries

First, we need to install the following libraries:

pandas
matplotlib
pywaffle
pypalettes
pyfonts
drawarrow
highlight_text

import pandas as pd
import matplotlib.pyplot as plt
from pywaffle import Waffle
from pypalettes import load_cmap
from pyfonts import load_font
from drawarrow import ax_arrow, fig_arrow
from highlight_text import fig_text

Dataset

First of all, we need to fetch the dataset from the tidytuesday repository, and clean it a bit:

url = "https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-03-04/longbeach.csv"
df = pd.read_csv(url)
df["outcome_year"] = pd.to_datetime(df["outcome_date"]).dt.year
df = df[df["outcome_type"] == "adoption"]
df = df[["outcome_year", "animal_type"]]
df_agg = df.groupby(["outcome_year", "animal_type"], as_index=False).size()
df_agg = df_agg[~df_agg["animal_type"].isin(["amphibian", "livestock"])]
all_years = df_agg["outcome_year"].unique()
all_animals = df_agg["animal_type"].unique()
idx = pd.MultiIndex.from_product(
    [all_years, all_animals], names=["outcome_year", "animal_type"]
)
df_agg = (
    df_agg.set_index(["outcome_year", "animal_type"])
    .reindex(idx, fill_value=0)
    .reset_index()
)
df_agg = df_agg.sort_values(["outcome_year", "animal_type"])
df_agg[df_agg["outcome_year"] == 2024].sort_values("size")
df_agg["iscatordog"] = df_agg["animal_type"].isin(["cat", "dog"])
df_agg.head()

	outcome_year	animal_type	size	iscatordog
0	2017.0	bird	1	False
1	2017.0	cat	222	True
2	2017.0	dog	205	True
3	2017.0	guinea pig	3	False
4	2017.0	other	1	False

Minimalist version

The core of the chart is to create a Figure with 8 Axes (one for each year).

Then, we iterate each of those Axes and add the waffle chart for this year.

The main trick here is to get the total value with the year with highest sum (max_year_value), and use it to fill all other years with white squares to ensure waffle charts have different sizes. Otherwise they'll have the same size and don't really make sense.

n_colors = df_agg["animal_type"].nunique()
colors = load_cmap("CrystalGems", keep_first_n=n_colors, shuffle=3).colors + ["white"]

max_year_value = df_agg[df_agg["outcome_year"] == 2024]["size"].sum()

ncols = df_agg["outcome_year"].nunique()
fig, axs = plt.subplots(ncols=ncols, figsize=(16, 10))

for year, ax in zip(df_agg["outcome_year"].unique(), axs):
    values = list(df_agg[df_agg["outcome_year"] == year]["size"].values)
    values = sorted(values, reverse=True)
    values.append(max_year_value - sum(values))

    Waffle.make_waffle(
        ax=ax,
        rows=50,
        columns=10,
        values=values,
        vertical=True,
        colors=colors,
    )

    ax.text(x=0.1, y=-0.04, s=f"{int(year)}", fontsize=14, ha="center")

plt.show()

Labels

All labels are added using the text() function, with trials and errors to find the perfect location. We also use the associate colors to serve as a legend.

The process is pretty much the same for arrows, but using ax_arrow()/fig_arrow().

n_colors = df_agg["animal_type"].nunique()
colors = load_cmap("CrystalGems", keep_first_n=n_colors, shuffle=3).colors + ["white"]

max_year_value = df_agg[df_agg["outcome_year"] == 2024]["size"].sum()

ncols = df_agg["outcome_year"].nunique()
fig, axs = plt.subplots(ncols=ncols, figsize=(16, 10))

for year, ax in zip(df_agg["outcome_year"].unique(), axs):
    values = list(df_agg[df_agg["outcome_year"] == year]["size"].values)
    values = sorted(values, reverse=True)
    values.append(max_year_value - sum(values))

    Waffle.make_waffle(
        ax=ax,
        rows=50,
        columns=10,
        values=values,
        vertical=True,
        colors=colors,
    )

    ax.text(x=0.1, y=-0.04, s=f"{int(year)}", fontsize=14, ha="center")

label_params = dict(
    x=1.05,
    size=15,
    transform=axs[ncols - 1].transAxes,
    clip_on=False,
)
axs[ncols - 1].text(y=0.24, s="Cat", color=colors[0], **label_params)
axs[ncols - 1].text(y=0.75, s="Dog", color=colors[1], **label_params)
axs[ncols - 1].text(y=0.96, s="Rabbit", color=colors[2], **label_params)
axs[ncols - 1].text(y=1.02, s="Guinea Pig", color=colors[3], **label_params)
axs[ncols - 1].text(y=0.99, s="Bird", color=colors[4], **label_params)
fig.text(
    x=0.56,
    y=0.38,
    s="Reptile",
    size=15,
    color=colors[5],
)
fig.text(
    x=0.47,
    y=0.45,
    s="Other",
    size=15,
    color=colors[6],
)

arrow_params = dict(
    ax=axs[ncols - 1],
    transform=axs[ncols - 1].transAxes,
    clip_on=False,
)
ax_arrow([1.04, 1.03], [0.6, 1.01], color=colors[3], radius=0.5, **arrow_params)
fig_arrow(
    [0.57, 0.37], [0.558, 0.31], color=colors[5], radius=0.1, clip_on=False, fig=fig
)
fig_arrow(
    [0.48, 0.44], [0.45, 0.38], color=colors[6], radius=-0.3, clip_on=False, fig=fig
)

plt.show()

y-scale

Due to the way PyWaffle works, we have to create the yscale line ourselves.

We use the text() function for labels, and axhline() avhline() for the lines.

n_colors = df_agg["animal_type"].nunique()
colors = load_cmap("CrystalGems", keep_first_n=n_colors, shuffle=3).colors + ["white"]

max_year_value = df_agg[df_agg["outcome_year"] == 2024]["size"].sum()

ncols = df_agg["outcome_year"].nunique()
fig, axs = plt.subplots(ncols=ncols, figsize=(16, 10))

# Add y scale
y_ticks = [0, 300, 600, 900, 1200, 1500]
for y_tick in y_ticks:
    axs[0].text(
        x=-0.25,
        y=y_tick / max_year_value,
        s=f"{y_tick}",
        size=12,
        va="center",
        ha="right",
        transform=axs[0].transAxes,
    )
    axs[0].axhline(
        y=y_tick / max_year_value,
        xmin=-0.2,
        xmax=-0.1,
        clip_on=False,
        color="black",
        linewidth=0.7,
    )
axs[0].axvline(x=-0.03, ymax=1.05, clip_on=False, color="black", linewidth=0.7)

for year, ax in zip(df_agg["outcome_year"].unique(), axs):
    values = list(df_agg[df_agg["outcome_year"] == year]["size"].values)
    values = sorted(values, reverse=True)
    values.append(max_year_value - sum(values))

    Waffle.make_waffle(
        ax=ax,
        rows=50,
        columns=10,
        values=values,
        vertical=True,
        colors=colors,
    )

    ax.text(x=0.1, y=-0.04, s=f"{int(year)}", fontsize=14, ha="center")

label_params = dict(
    x=1.05,
    size=15,
    transform=axs[ncols - 1].transAxes,
    clip_on=False,
)
axs[ncols - 1].text(y=0.24, s="Cat", color=colors[0], **label_params)
axs[ncols - 1].text(y=0.75, s="Dog", color=colors[1], **label_params)
axs[ncols - 1].text(y=0.96, s="Rabbit", color=colors[2], **label_params)
axs[ncols - 1].text(y=1.02, s="Guinea Pig", color=colors[3], **label_params)
axs[ncols - 1].text(y=0.99, s="Bird", color=colors[4], **label_params)
fig.text(
    x=0.56,
    y=0.38,
    s="Reptile",
    size=15,
    color=colors[5],
)
fig.text(
    x=0.47,
    y=0.45,
    s="Other",
    size=15,
    color=colors[6],
)

arrow_params = dict(
    ax=axs[ncols - 1],
    transform=axs[ncols - 1].transAxes,
    clip_on=False,
)
ax_arrow([1.04, 1.03], [0.6, 1.01], color=colors[3], radius=0.5, **arrow_params)
fig_arrow(
    [0.57, 0.37], [0.558, 0.31], color=colors[5], radius=0.1, clip_on=False, fig=fig
)
fig_arrow(
    [0.48, 0.44], [0.45, 0.38], color=colors[6], radius=-0.3, clip_on=False, fig=fig
)

plt.show()

Title and fonts

We use 3 variations of the Roboto font here, depending on the labels. We load them using load_font().

The title/subtitle and the captions are added via text().

repo_url = "https://github.com/googlefonts/roboto-2/blob/main/src/hinted"
lightfont = load_font(f"{repo_url}/Roboto-Light.ttf?raw=true")
regularfont = load_font(f"{repo_url}/Roboto-Regular.ttf?raw=true")
italicfont = load_font(f"{repo_url}/Roboto-BlackItalic.ttf?raw=true")

n_colors = df_agg["animal_type"].nunique()
colors = load_cmap("CrystalGems", keep_first_n=n_colors, shuffle=3).colors + ["white"]

max_year_value = df_agg[df_agg["outcome_year"] == 2024]["size"].sum()

ncols = df_agg["outcome_year"].nunique()
fig, axs = plt.subplots(ncols=ncols, figsize=(16, 10))

y_ticks = [0, 300, 600, 900, 1200, 1500]
for y_tick in y_ticks:
    axs[0].text(
        x=-0.25,
        y=y_tick / max_year_value,
        s=f"{y_tick}",
        size=12,
        va="center",
        ha="right",
        transform=axs[0].transAxes,
        font=lightfont,
    )
    axs[0].axhline(
        y=y_tick / max_year_value,
        xmin=-0.2,
        xmax=-0.1,
        clip_on=False,
        color="black",
        linewidth=0.7,
    )
axs[0].axvline(x=-0.03, ymax=1.05, clip_on=False, color="black", linewidth=0.7)


for year, ax in zip(df_agg["outcome_year"].unique(), axs):
    values = list(df_agg[df_agg["outcome_year"] == year]["size"].values)
    values = sorted(values, reverse=True)
    values.append(max_year_value - sum(values))

    Waffle.make_waffle(
        ax=ax,
        rows=50,
        columns=10,
        values=values,
        vertical=True,
        colors=colors,
    )

    ax.text(x=0.1, y=-0.04, s=f"{int(year)}", fontsize=14, ha="center", font=lightfont)
    adj_y = 0.03
    ax.text(
        x=0.03,
        y=sum(values[:-1]) / max_year_value + adj_y,
        s=f"{sum(values[:-1])}",
        fontsize=15,
        ha="center",
        font=italicfont,
    )

label_params = dict(
    x=1.05,
    size=15,
    transform=axs[ncols - 1].transAxes,
    font=regularfont,
    clip_on=False,
)
axs[ncols - 1].text(y=0.24, s="Cat", color=colors[0], **label_params)
axs[ncols - 1].text(y=0.75, s="Dog", color=colors[1], **label_params)
axs[ncols - 1].text(y=0.96, s="Rabbit", color=colors[2], **label_params)
axs[ncols - 1].text(y=1.02, s="Guinea Pig", color=colors[3], **label_params)
axs[ncols - 1].text(y=0.99, s="Bird", color=colors[4], **label_params)
fig.text(
    x=0.56,
    y=0.38,
    s="Reptile",
    size=15,
    font=regularfont,
    color=colors[5],
)
fig.text(
    x=0.47,
    y=0.45,
    s="Other",
    size=15,
    font=regularfont,
    color=colors[6],
)

arrow_params = dict(
    ax=axs[ncols - 1],
    transform=axs[ncols - 1].transAxes,
    clip_on=False,
)
ax_arrow([1.04, 1.03], [0.6, 1.01], color=colors[3], radius=0.5, **arrow_params)
fig_arrow(
    [0.57, 0.37], [0.558, 0.31], color=colors[5], radius=0.1, clip_on=False, fig=fig
)
fig_arrow(
    [0.48, 0.44], [0.45, 0.38], color=colors[6], radius=-0.3, clip_on=False, fig=fig
)

title = """
Number of animals adopted at the\nLong Beach animal shelter
"""
fig.text(x=0.14, y=0.9, s=title, size=30, font=lightfont, va="top")

share_catdog = df_agg[df_agg["iscatordog"]]["size"].sum() / df_agg["size"].sum() * 100
subtitle = f"""
<Dogs> and <Cats> represent {share_catdog:.0f}% of all adoptions during the period
"""
fig_text(
    x=0.14,
    y=0.75,
    s=subtitle,
    size=18,
    font=lightfont,
    color="#afafaf",
    va="top",
    highlight_textprops=[{"color": colors[1]}, {"color": colors[0]}],
)

source_params = dict(ha="right", font=lightfont, color="#9a9a9a", size=10)
fig.text(x=0.9, y=0.04, s="TidyTuesday - 2025-03-04", **source_params)
fig.text(x=0.9, y=0.02, s="Viz: Joseph Barbier", **source_params)

fig.savefig(
    "../../static/graph/web-waffle-chart-with-groups-evolution.png",
    dpi=300,
    bbox_inches="tight",
)

plt.show()

Going further

Treemap

Venn Diagram

Donut

Pie Chart

Dendrogram

Circular Packing

Waffle

🚨 Grab the Data To Viz poster!

Do you know all the chart types? Do you know which one you should pick? I made a decision tree that answers those questions. You can download it for free!