How to Create Multiple Bar Charts in Matplotlib (Grouped, Side by Side)

To create a multiple bar chart in Matplotlib (also called a grouped, double or side-by-side bar chart), draw each series with ax.bar() at positions shifted by the bar width, for example x - width/2 and x + width/2, then put the group labels in the middle with ax.set_xticks(x, labels). With pandas, df.plot(kind="bar") draws a grouped bar for every column. This guide covers both, plus any number of series, horizontal bars, and several separate bar charts side by side.

Tested with Python 3.12.5, Matplotlib 3.11.2, NumPy 2.5.3 and pandas 3.0.6; the charts are real Matplotlib windows. Reference: Matplotlib grouped bar chart example and DataFrame.plot.bar.

Double bar graph (two bars side by side)

import matplotlib.pyplot as plt
import numpy as np

quarters = ["Q1", "Q2", "Q3", "Q4"]
sales_2024 = [120, 135, 150, 170]
sales_2025 = [128, 149, 162, 195]

x = np.arange(len(quarters))        # 0, 1, 2, 3: one group per quarter
width = 0.38                        # width of each bar

fig, ax = plt.subplots(figsize=(7, 4.5))
bars1 = ax.bar(x - width / 2, sales_2024, width, label="2024")   # shift left
bars2 = ax.bar(x + width / 2, sales_2025, width, label="2025")   # shift right
ax.bar_label(bars1, padding=2)
ax.bar_label(bars2, padding=2)

ax.set_xticks(x, quarters)          # labels in the middle of each group
ax.set_ylabel("Sales (thousand USD)")
ax.set_title("Double bar graph: sales by quarter")
ax.legend()
plt.tight_layout()
plt.show()

The trick is the x positions: np.arange() gives one position per group, and each series is moved half a bar width to the left or right so the bars touch instead of overlapping.

Any number of bars per group

For n series, make each bar 0.8 / n wide and move series i by (i - (n - 1) / 2) * width. This prints the positions it produces:

import numpy as np

groups = 4                     # quarters
series = ["2023", "2024", "2025"]
width = 0.8 / len(series)      # the bars of one group fill 80% of the space between groups

x = np.arange(groups)
for i, name in enumerate(series):
    offset = (i - (len(series) - 1) / 2) * width
    print(f"{name}: offset {offset:+.3f} -> bar centres {np.round(x + offset, 3)}")

Output:

2023: offset -0.267 -> bar centres [-0.267  0.733  1.733  2.733]
2024: offset +0.000 -> bar centres [0. 1. 2. 3.]
2025: offset +0.267 -> bar centres [0.267 1.267 2.267 3.267]

import matplotlib.pyplot as plt
import numpy as np

products = ["Laptops", "Phones", "Tablets", "Watches", "Audio"]
regions = {"North": [42, 55, 20, 18, 25], "South": [38, 61, 24, 15, 30],
           "East": [50, 47, 19, 22, 28], "West": [45, 52, 27, 20, 21]}

x = np.arange(len(products))
width = 0.8 / len(regions)

fig, ax = plt.subplots(figsize=(9, 4.5))
for i, (region, values) in enumerate(regions.items()):
    offset = (i - (len(regions) - 1) / 2) * width
    ax.bar(x + offset, values, width, label=region)

ax.set_xticks(x, products)
ax.set_ylabel("Units sold (thousands)")
ax.legend(title="Region", ncols=4, loc="upper right")
plt.tight_layout()
plt.show()

Bar plot of multiple columns with pandas

If your data is in a DataFrame, each column becomes a series and each row a group. No offset calculation is needed:

import matplotlib.pyplot as plt
import pandas as pd

df = pd.DataFrame({"2023": [310, 280, 190], "2024": [335, 300, 210], "2025": [360, 290, 240]},
                  index=["Electronics", "Clothing", "Groceries"])

ax = df.plot(kind="bar", figsize=(7, 4.5), rot=0, width=0.8)   # one bar per column, grouped by row
ax.set_ylabel("Revenue (thousand USD)")
ax.set_title("pandas: bar plot of multiple columns")
for container in ax.containers:
    ax.bar_label(container, fontsize=8)
plt.tight_layout()
plt.show()

Use kind="barh" for horizontal bars and stacked=True for stacked bars. To choose columns, pass y=["col1", "col2"].

Horizontal grouped bar chart

import matplotlib.pyplot as plt
import numpy as np

skills = ["Python", "SQL", "Excel", "Tableau"]
junior = [60, 55, 80, 30]
senior = [90, 85, 70, 65]

y = np.arange(len(skills))
height = 0.38
fig, ax = plt.subplots(figsize=(7, 4))
ax.barh(y - height / 2, junior, height, label="Junior")
ax.barh(y + height / 2, senior, height, label="Senior")
ax.set_yticks(y, skills)
ax.invert_yaxis()
ax.set_xlabel("Average score")
ax.legend()
plt.tight_layout()
plt.show()

Several separate bar charts side by side

Sometimes “multiple bar charts” means several charts, not grouped bars. Put one bar chart in each subplot and share the y-axis so they are easy to compare:

import matplotlib.pyplot as plt

data = {"Chicago": [5, 7, 12, 18, 23, 27], "Miami": [24, 25, 26, 28, 30, 31],
        "Seattle": [7, 8, 10, 13, 16, 19]}
months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun"]

fig, axes = plt.subplots(1, 3, figsize=(11, 3.6), sharey=True)
for ax, (city, temps), color in zip(axes, data.items(), ["tab:blue", "tab:orange", "tab:green"]):
    ax.bar(months, temps, color=color)       # one separate bar chart per subplot
    ax.set_title(city)
axes[0].set_ylabel("Temperature (°C)")
plt.tight_layout()
plt.show()

Grouped or stacked?

Chart	Best for
Grouped (side by side)	Comparing the series with each other within each group
Stacked	Showing the total of each group and how it is made up
Separate subplots	Many groups or series with different ranges

See how to create a stacked bar chart in Matplotlib for the stacked version.

More Matplotlib bar chart tutorials:

• Create a stacked bar chart in Matplotlib
• Plot a bar chart in Matplotlib
• Show values on bar charts
• Add error bars to a bar chart

Frequently asked questions