5 Best Ways to Use Python Pandas to Draw a Bar Plot with Median as the Central Tendency Estimate

by

💡 Problem Formulation: In data visualization, a bar plot is a common tool to represent the distribution of data. While the mean is the default estimate of central tendency, sometimes the median is more appropriate due to its robustness to outliers. In Python’s Pandas library, creating a bar plot with the median as the central tendency involves custom configuration. This article aims to demonstrate how to plot such bar charts, detailing methods to derive and visualize the median values within a dataset. The input is a Pandas DataFrame and the desired output is a bar plot with bars representing the median values of the chosen data columns.

Method 1: Basic Bar Plot with Median Calculation

This method involves calculating the median of each column in the DataFrame and then creating a bar plot from these median values. This is the most straightforward way to create a bar plot with median as the estimator.

Here’s an example:

import pandas as pd
import matplotlib.pyplot as plt

# Sample DataFrame
data = {'A': [12, 4, 5, 44, 1], 'B': [5, 2, 54, 3, 15], 'C': [20, 16, 7, 3, 8]}
df = pd.DataFrame(data)

# Calculate medians
medians = df.median()

# Plot
medians.plot(kind='bar')
plt.show()

The output is a bar plot with bars representing the median values for columns A, B, and C.

This code first creates a DataFrame from a dictionary of lists. It then calculates the median of each column using Pandas’ median() function and creates a bar plot from these values. The plot() method on the DataFrame is used with the kind parameter set to ‘bar’ to create the bar plot.

Method 2: Aggregate with Custom Function

One versatile approach involves using the agg() method. This method allows for the application of a custom function across the DataFrame, empowering you to calculate the median directly during the aggregation phase for the plot.

Here’s an example:

import pandas as pd
import matplotlib.pyplot as plt

# Sample DataFrame
data = {'Scores': [12, 20, 5, 8, 15], 'Category': ['A', 'A', 'B', 'B', 'A']}
df = pd.DataFrame(data)

# Plot with custom aggregation
df.groupby('Category').agg('median').plot(kind='bar')
plt.show()

The output is a bar plot with bars representing the median scores for each category.

Here, the groupby() method is leveraged to aggregate data based on category. After the grouping, we apply the agg() function with ‘median’ passed to it for calculating the median scores. A bar plot is then drawn using the aggregated data.

Method 3: Using Seaborn’s barplot Function

The library Seaborn, a statistical data visualization library based on Matplotlib, provides a high-level barplot function that can take the estimator as an argument.

Here’s an example:

import pandas as pd
import seaborn as sns

# Sample DataFrame
data = {'Values': [10, 14, 17, 13, 12, 16], 'Groups': ['X', 'X', 'Y', 'Y', 'Z', 'Z']}
df = pd.DataFrame(data)

# Draw a barplot with median
sns.barplot(x='Groups', y='Values', data=df, estimator=np.median)
plt.show()

The output is a bar plot with bars representing the median values for each group.

This example uses seaborn’s barplot() function, which has an estimator parameter that we set to np.median for calculating the median. Since Seaborn operates on Pandas DataFrames, you can pass your data directly into the function.

Method 4: Pivot Table with a Bar Plot

Creating a pivot table can be a powerful method for reshaping the data to calculate the median per category before plotting. This method gives you more control over the data manipulation process.

Here’s an example:

import pandas as pd
import matplotlib.pyplot as plt

# Sample data
data = {'Product': ['A', 'A', 'B', 'B', 'C', 'C'], 'Sales': [100, 230, 150, 300, 400, 320]}
df = pd.DataFrame(data)

# Pivot table
pivot_table = df.pivot_table(index='Product', values='Sales', aggfunc='median')

# Bar plot
pivot_table.plot(kind='bar')
plt.show()

The output is a bar plot with bars representing the median sales for each product.

The code snippet first creates a pivot table from the DataFrame using the pivot_table method specifying ‘median’ as the aggregation function. We then plot this pivot table, which simplifies the visualization of median values for each product.

Bonus One-Liner Method 5: Direct Median Plotting

For a quick and effective one-liner, you can directly plot the median of the DataFrame without prior calculation. This method is most suitable for quick visual checks when writing exploratory code.

Here’s an example:

import pandas as pd

# Sample DataFrame
df = pd.DataFrame({'X': [1, 2, 3, 11], 'Y': [4, 5, 2, 8], 'Z': [1, 0, 3, 9]})

# Direct plotting
df.median().plot(kind='bar')

The output is a bar plot with bars representing the median values for columns X, Y, and Z.

With this concise line, we utilize the chainability of methods in Pandas by directly calling median() followed by plot().

Summary/Discussion

  • Method 1: Basic Bar Plot with Median Calculation. Easy to understand and implement. May require additional steps for complex data manipulation.
  • Method 2: Aggregate with Custom Function. Allows more complex aggregation logic. Involves more Pandas operations.
  • Method 3: Using Seaborn’s barplot Function. Provides an elegant and concise approach for statistical plots. Requires Seaborn, which might not be installed by default.
  • Method 4: Pivot Table with a Bar Plot. Great for multi-category median analysis. Can be verbose for simple tasks.
  • Method 5: Direct Median Plotting. Quickest one-liner solution. Offers less flexibility for customization.