Grouped barplot with the total represented in the background

In this post, we'll see how to create a grouped barplot with the total of each subgroup in the background and how to customize it to make it look nice

The data

Let's get started by importing libraries and loading the data

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

data = pd.DataFrame(
    {
        "value": [364, 232, 143, 357, 204, 131, 254, 158, 91, 196, 122, 74],
        "category": [
            "West",
            "West",
            "West",
            "East",
            "East",
            "East",
            "Central",
            "Central",
            "Central",
            "South",
            "South",
            "South",
        ],
        "sub_category": [
            "Consumer",
            "Corporate",
            "Home office",
            "Consumer",
            "Corporate",
            "Home office",
            "Consumer",
            "Corporate",
            "Home office",
            "Consumer",
            "Corporate",
            "Home office",
        ],
    }
)

cat_order = ["West", "East", "Central", "South"]
data["category"] = pd.Categorical(data["category"], categories=cat_order, ordered=True)
data = data.sort_values(["category", "value"], ascending=[True, False])

The dataset looks like this:

data

	value	category	sub_category
0	364	West	Consumer
1	232	West	Corporate
2	143	West	Home office
3	357	East	Consumer
4	204	East	Corporate
5	131	East	Home office
6	254	Central	Consumer
7	158	Central	Corporate
8	91	Central	Home office
9	196	South	Consumer
10	122	South	Corporate
11	74	South	Home office

Default grouped barplot

pivot_df reshapes the data so that each row is a category and each column is a sub-category.
x gives the base x-positions for the categories.
The for loop goes through each sub-category and offsets its bars horizontally using i * bar_width so they appear grouped.
Each ax.bar() call draws one set of bars (for one sub-category) at the correct shifted position.

pivot_df = data.pivot(index="category", columns="sub_category", values="value")

fig, ax = plt.subplots(figsize=(7, 5))

bar_width = 0.25
x = np.arange(len(pivot_df.index))
for i, sub_cat in enumerate(pivot_df.columns):
    ax.bar(x + i * bar_width, pivot_df[sub_cat], width=bar_width, label=sub_cat)

ax.set_xlabel("Category")
ax.set_ylabel("Value")
ax.set_title("Regions - Revenue by Segment", loc="left")
ax.set_xticks(x + bar_width / 2)
ax.set_xticklabels(pivot_df.index)
ax.legend(title="Sub Category")

plt.show()

Total values in the background

This version adds total bars behind the grouped bars.

It first calculates totals for each category (sum across sub-categories).
ax.bar(x, totals, ...) draws wide, light grey bars as a background, showing the total per category.
The for loop then overlays the narrower, colored bars for each sub-category, centered around each category position.

So while both use a loop to draw grouped bars, this one also visualizes the total, making it easier to compare individual sub-category values to the whole. The offset (i - 0.5) * bar_width ensures the grouped bars are centered over the total bar.

pivot_df = data.pivot(index="category", columns="sub_category", values="value")
totals = pivot_df.sum(axis=1)

fig, ax = plt.subplots(figsize=(7, 5))

bar_width = 0.25
x = np.arange(len(pivot_df.index))

ax.bar(x, totals, width=bar_width * 3.5, color="lightgrey", zorder=0, label="Total")

for i, sub_cat in enumerate(pivot_df.columns):
    ax.bar(
        x + (i - 1) * bar_width,
        pivot_df[sub_cat],
        width=bar_width,
        label=sub_cat,
        zorder=1,
    )

ax.set_xlabel("Category")
ax.set_ylabel("Value")
ax.set_title("Regions - Revenue by Segment", loc="left")
ax.set_xticks(x)
ax.set_xticklabels(pivot_df.index)
ax.legend(title="Sub Category")

fig.savefig(
    "../../static/graph/grouped-barplot-with-the-total-of-each-group-represented-as-a-grey-rectangle.png",
    dpi=100,
    bbox_inches="tight",
)
plt.show()

Customization

Let's make it look nicer!

from highlight_text import fig_text

pivot_df = data.pivot(index="category", columns="sub_category", values="value")
totals = pivot_df.sum(axis=1)

fig, ax = plt.subplots(figsize=(7, 5))

bar_width = 0.25
x = np.arange(len(pivot_df.index))
colors = ["#E0AC2B", "#E85252", "#6689C6"]
x = np.arange(len(pivot_df.index))

ax.bar(x, totals, width=bar_width * 3.5, color="#E0E0E0", zorder=0)

for i, sub_cat in enumerate(pivot_df.columns):
    ax.bar(
        x + (i - 1) * bar_width,
        pivot_df[sub_cat],
        width=bar_width,
        zorder=1,
        color=colors[i],
    )

ax.set_xticks(x)
ax.set_xticklabels(pivot_df.index)
ax.set_yticks(list(range(0, 801, 100)))
ax.spines[["top", "right", "bottom"]].set_visible(False)
ax.tick_params(axis="x", size=0, pad=10)
ax.tick_params(axis="y", labelsize=8)

fig.text(x=0.12, y=1, s="Regions - Revenue by Segment", size=14)
fig_text(
    x=0.12,
    y=0.97,
    s="<Consumer> <Corporate Home> <Office>",
    highlight_textprops=[
        {"color": colors[0]},
        {"color": colors[1]},
        {"color": colors[2]},
    ],
)

fig.savefig(
    "../../static/graph/grouped-barplot-with-the-total-of-each-group-represented-as-a-grey-rectangle-2.png",
    dpi=100,
    bbox_inches="tight",
)
plt.show()

Label values

The very last step in improving this graph is to add the values for each sub-category and each overall category. The key is to use for loops to dynamically add these labels.

pivot_df = data.pivot(index="category", columns="sub_category", values="value")
totals = pivot_df.sum(axis=1)

fig, ax = plt.subplots(figsize=(7, 5))

bar_width = 0.25
x = np.arange(len(pivot_df.index))
colors = ["#E0AC2B", "#E85252", "#6689C6"]

ax.bar(x, totals, width=bar_width * 3.5, color="#E0E0E0", zorder=0)

for i, sub_cat in enumerate(pivot_df.columns):
    values = pivot_df[sub_cat]
    ax.bar(
        x + (i - 1) * bar_width,
        values,
        width=bar_width,
        zorder=1,
        color=colors[i],
    )
    for j, val in enumerate(values):
        ax.text(
            x[j] + (i - 1) * bar_width,
            val + 5,
            f"{val:.0f}",
            ha="center",
            va="bottom",
            fontsize=8,
        )

for i, total in enumerate(totals):
    ax.text(
        x[i],
        total + 10,
        f"{total:.0f}",
        ha="center",
        va="bottom",
        fontsize=9,
        color="#555555",
    )

ax.set_xticks(x)
ax.set_xticklabels(pivot_df.index)
ax.set_yticks(list(range(0, 801, 100)))
ax.spines[["top", "right", "bottom"]].set_visible(False)
ax.tick_params(axis="x", size=0, pad=10)
ax.tick_params(axis="y", labelsize=8)

fig.text(x=0.12, y=1, s="Regions - Revenue by Segment", size=14)
fig_text(
    x=0.12,
    y=0.97,
    s="<Consumer> <Corporate Home> <Office>",
    highlight_textprops=[
        {"color": colors[0]},
        {"color": colors[1]},
        {"color": colors[2]},
    ],
)

fig.savefig(
    "../../static/graph/grouped-barplot-with-the-total-of-each-group-represented-as-a-grey-rectangle-3.png",
    dpi=100,
    bbox_inches="tight",
)
plt.show()

Going further

Learn more about barplots
How to create a stacked barplot
How to create a percent stacked barplot

Barplot

Spider / Radar

Wordcloud

Parallel

Lollipop

Circular Barplot

Table

🚨 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!