barinbar

A Python package for creating bar-in-bar charts and treemap bar charts using Matplotlib.

Inspiration

This package was created in response to discussions and implementations of bar-in-bar charts across different platforms:

• Original LinkedIn Post by Colin Tomb: Link
• Tableau Implementation (Kevin Flerlage, Sebastine Amede, Darragh Murray, Brittany Rosenau): Blog Post
• R Implementation (Brian Julius): Post
• Deneb Implementation (Daniel Marsh-Patrick): Blog Post

This package aims to provide straightforward Python (Matplotlib) alternatives.

Installation

pip install barinbar

(Note: This assumes the package is published on PyPI.)

Development Setup (using Poetry)

If you want to contribute or run the code locally:

1. Install Poetry: Follow the instructions on the official Poetry website.
2. Clone the repository:

   git clone <your-repo-url>
   cd barinbar
   

3. Install dependencies: This command creates a virtual environment and installs all dependencies (including development tools).

   poetry install --with dev # Use --with dev instead of --all-extras for newer Poetry
   

4. Activate the virtual environment:

   poetry shell
   

5. Now you can run scripts or linters, e.g., python barinbar/plot.py.

Requirements

• Python >= 3.10
• matplotlib
• pandas
• numpy
• squarify

Usage

Bar-in-Bar Chart

The main function is plot_bar_in_bar. It requires a pandas DataFrame structured with columns for the main category, sub-category, segment value, and total category value.

import pandas as pd
from barinbar import plot_bar_in_bar, BarinbarStyle, plot_treemap_bars # Import from package root
from matplotlib.cm import get_cmap
import numpy as np # Import numpy

# 1. Prepare your data
data = {
    "Region": ["West"]*3 + ["East"]*3 + ["Central"]*3 + ["South"]*3,
    "Segment": ["Consumer", "Corporate", "Home Office"] * 4,
    "Revenue": [364, 232, 143, 357, 204, 131, 254, 158, 91, 196, 122, 74],
}
revenues = pd.DataFrame(data)

# Calculate totals per main category
region_totals = revenues.groupby("Region")["Revenue"].sum().reset_index()
region_totals = region_totals.rename(columns={"Revenue": "TotalRevenue"})

# Merge totals back into the main DataFrame
revenues = pd.merge(revenues, region_totals, on="Region")

# Sorting is now handled internally by _prepare_data

# 2. Define Styling (Optional)
default_color_palette = get_cmap("Dark2")
custom_colors = {
    "Consumer": default_color_palette(0),
    "Corporate": default_color_palette(1),
    "Home Office": default_color_palette(2),
}
my_style = BarinbarStyle(
    fontname='serif',
    colors=custom_colors,
    figsize=(12, 7)
)

# 3. Plot the chart
# Example with total background width
print("Plotting Bar-in-Bar (Total Background)...")
plot_bar_in_bar(
    df=revenues.copy(),
    category_col="Region",
    subcategory_col="Segment",
    value_col="Revenue",
    total_col="TotalRevenue",
    title="Revenue by Region and Segment (Background=Total Width)",
    style=my_style,
    back_ground_width="total" # Specify background width option
)

# Example with single background width
print("Plotting Bar-in-Bar (Single Background)...")
plot_bar_in_bar(
    df=revenues.copy(),
    category_col="Region",
    subcategory_col="Segment",
    value_col="Revenue",
    total_col="TotalRevenue",
    title="Revenue by Region and Segment (Background=Single Width)",
    style=my_style,
    back_ground_width="single" # Specify background width option
)

Treemap Bars Chart

The plot_treemap_bars function generates horizontally arranged treemaps, where each treemap represents a group, and the rectangles within represent categories. The width of each treemap is consistent, while its height is proportional to the total value of the group. Groups are sorted left-to-right from highest total value to lowest.

# (Assuming 'revenues' DataFrame and 'my_style' from previous example)

print("Plotting Treemap Bars...")
plot_treemap_bars(
    df=revenues.copy(),
    category_col="Segment",  # Categories within each treemap box
    group_col="Region",      # How data is grouped into separate treemaps
    value_col="Revenue",     # Determines size and height scaling
    title="Regional Revenue Distribution by Segment (Treemap)",
    figsize=(14, 7),         # Adjust figsize as needed
    value_format_str="${value}K",
    fontname=my_style.fontname, # Reuse font from style
    base_palette="tab10",     # Choose a suitable palette
)

This will generate and display the respective charts.

Customization

You can customize the appearance of the bar-in-bar chart by creating an instance of the BarinbarStyle dataclass and modifying its attributes (e.g., figsize, colors, fontname, fontsize, background_color). Pass this style object to the plot_bar_in_bar function.

For plot_treemap_bars, customization is done via function arguments like figsize, value_format_str, fontname, and base_palette.

from barinbar import BarinbarStyle # Import from package root

custom_style = BarinbarStyle(
    background_color="#F0F0F0",
    figsize=(10, 6),
    fontsize=18,
    small_fontsize=11,
    fontname="Arial",
    colors={"Consumer": "blue", "Corporate": "green", "Home Office": "red"}
)

# ... then pass `style=custom_style` to the plot_bar_in_bar function

License

This project is licensed under the MIT License - see the LICENSE file for details.

Author

• Daniel Vecera - daniel.vecera@outlook.com