Create Stunning Dynamic Bar Chart Races in Python with Just 3 Lines of Code

Why dynamic bar charts?

Dynamic bar charts are one of the hottest forms of data visualisation, with many high‑view‑count videos on platforms like Bilibili. Existing tools such as Huaxian "Hanabi", Dijishu "DiShu Chart" and Flourish can create them, but Python developers often prefer a native solution.

Introducing bar_chart_race

The Bar Chart Race library is a powerful Python package that simplifies creating animated bar charts. It can be found on GitHub at https://github.com/dexplo/bar_chart_race with documentation at https://www.dexplo.org/bar_chart_race/ . Two versions exist (0.1 and 0.2); version 0.2 adds dynamic line charts and Plotly‑based bar charts.

Installation

Installation via pip install bar_chart_race only provides version 0.1, which lacks some features. In PyCharm’s Project Interpreter the same limitation appears. The workaround is to download the source from GitHub and install manually:

cd your_project_path/venv/lib/python3.7/site-packages/bar_chart_race-master
python setup.py install
# Installation success messages

After manual installation you can import the library normally.

Basic usage – generate a GIF in three lines

import bar_chart_race as bcr
# Load example dataset
df = bcr.load_dataset('covid19_tutorial')
# Create animated GIF
bcr.bar_chart_race(df, 'covid19_horiz.gif')

The resulting GIF demonstrates a simple dynamic bar chart.

Customising the chart

The library offers many parameters. Below are common examples:

Orientation : bcr.bar_chart_race(df, 'covid19_horiz.gif', orientation='v') Sorting (default descending, use sort='asc' for ascending)

Number of bars : n_bars=6 Fixed order : specify a list of categories to keep their order.

Fixed maximum axis : fixed_max=True Frames per period (controls smoothness): steps_per_period=3 Period length (ms) : period_length=200 Period summary : define a function returning a dict with text position and content.

Perpendicular bar : supply a function returning a quantile value.

Colour map : cmap='accent' or custom maps.

Remove duplicate colours : filter_column_colors=True Figure size & DPI : figsize=(5,3), dpi=100 Hide bar labels : label_bars=False Custom period label and title styling via dictionaries.

Text size : bar_label_size=4, tick_label_size=5, title_size='smaller' Global font settings for Chinese characters:

plt.rcParams['font.sans-serif'] = ['SimHei']  # Windows
plt.rcParams['font.sans-serif'] = ['Hiragino Sans GB']  # macOS
plt.rcParams['axes.unicode_minus'] = False

After setting the font, Chinese titles and labels render correctly.

Custom colour maps

Define a new colour list in _colormaps.py:

colormaps = {
    "new_colors": [
        '#ff812c', '#ff5a5a', '#00c5d2', '#a64dff', '#4e70f0', '#f95dba', '#ffce2b'
    ]
}

Then use it with cmap='new_colors' to apply the custom palette.

Real‑world example with custom data

Load a CSV containing Baidu Index data for characters from the TV series “Yu Huan Shui”, pivot it into the required format, and visualise it:

import pandas as pd
import bar_chart_race as bcr

df = pd.read_csv('yuhuanshui.csv', encoding='utf-8', header=0,
                 names=['name','number','day'])
df_result = pd.pivot_table(df, values='number', index='day',
                          columns='name', fill_value=0)
bcr.bar_chart_race(df_result, 'heat.gif', title='我是余欢水演职人员热度排行')

The final animated chart displays the popularity ranking over time.

Conclusion

The bar_chart_race library provides a concise, flexible way to create animated bar‑chart‑races in Python. With a handful of parameters you can control orientation, sorting, colour schemes, label styles, Chinese text support, and export formats (GIF, MP4). Exploring the source code reveals even more fine‑tuned options.