Plotly Data Visualization Tutorial

This document explains the Plotly customer churn analysis example provided in plotly_example.py.

Overview

The example demonstrates how to:

1. Generate simulated customer churn data
2. Create a DataFrame using pandas
3. Create multiple types of interactive plots using Plotly
4. Combine plots into a dashboard layout
5. Save the plot as an interactive HTML file
6. Display the plot in a web browser
7. Calculate and print basic churn statistics

Code Explanation

Imports and Setup

import os
import plotly.graph_objects as go
import plotly.express as px
from plotly.subplots import make_subplots
import pandas as pd
import numpy as np

output_dir = os.path.dirname(os.path.abspath(__file__))

This section imports necessary modules and sets up the output directory.

Data Generation

np.random.seed(42)
n_customers = 1000

data = {
    'CustomerID': range(1, n_customers + 1),
    'Churn': np.random.choice([0, 1], n_customers, p=[0.8, 0.2]),
    'Tenure': np.random.randint(1, 72, n_customers),
    'MonthlyCharges': np.random.uniform(20, 100, n_customers),
    'TotalCharges': np.random.uniform(100, 5000, n_customers),
    'Contract': np.random.choice(['Month-to-month', 'One year', 'Two year'], n_customers),
    'InternetService': np.random.choice(['DSL', 'Fiber optic', 'No'], n_customers),
    'CustomerService': np.random.randint(1, 6, n_customers)
}

df = pd.DataFrame(data)

Here, we generate simulated customer churn data and create a pandas DataFrame.

Creating Subplots

fig = make_subplots(
    rows=2, cols=2,
    specs=[[{'type': 'domain'}, {'type': 'xy'}],
           [{'type': 'xy'}, {'type': 'xy'}]],
    subplot_titles=('Churn Rate', 'Monthly Charges vs Tenure',
                    'Internet Service Distribution', 'Customer Service Ratings')
)

This creates a 2x2 grid for our four plots, specifying the appropriate type for each subplot. The 'domain' type is used for the pie chart, while 'xy' is used for the other plots that require x and y axes.

Plotting

# Plot 1: Churn Rate (Pie Chart)
churn_counts = df['Churn'].value_counts()
fig.add_trace(
    go.Pie(labels=['Retained', 'Churned'], values=churn_counts.values, hole=.3),
    row=1, col=1
)

# Plot 2: Monthly Charges vs Tenure (Scatter Plot)
fig.add_trace(
    go.Scatter(x=df['Tenure'], y=df['MonthlyCharges'], mode='markers',
               marker=dict(color=df['Churn'], colorscale='Viridis'),
               text=df['CustomerID'], hoverinfo='text+x+y'),
    row=1, col=2
)

# Plot 3: Internet Service Distribution (Bar Chart)
internet_service_counts = df['InternetService'].value_counts()
fig.add_trace(
    go.Bar(x=internet_service_counts.index, y=internet_service_counts.values),
    row=2, col=1
)

# Plot 4: Customer Service Ratings (Histogram)
fig.add_trace(
    go.Histogram(x=df['CustomerService'], nbinsx=5),
    row=2, col=2
)

This section creates four different types of plots using Plotly: a pie chart, a scatter plot, a bar chart, and a histogram.

Updating Layout and Saving the Plot

fig.update_layout(height=800, title_text="Customer Churn Analysis Dashboard")

output_file_path = os.path.join(output_dir, 'churn_analysis.html')
fig.write_html(output_file_path)

fig.show()

This updates the layout, saves the plot as an interactive HTML file, and displays it in a web browser.

Statistical Analysis

churn_rate = df['Churn'].mean() * 100
avg_tenure = df['Tenure'].mean()
avg_monthly_charges = df['MonthlyCharges'].mean()
most_common_contract = df['Contract'].mode().iloc[0]

print(f"Churn Rate: {churn_rate:.2f}%")
print(f"Average Tenure: {avg_tenure:.2f} months")
print(f"Average Monthly Charges: ${avg_monthly_charges:.2f}")
print(f"Most Common Contract Type: {most_common_contract}")

This section calculates and prints basic statistics about the churn data.

Running the Example

To run this example:

1. Ensure you have Plotly, pandas, and numpy installed:

   pip install plotly pandas numpy
   

2. Run the script:

   python plotly_example.py
   

The script will generate an interactive HTML file 'churn_analysis.html' in the same directory as the script, open the plot in your default web browser, and print some basic statistics about the customer churn data.

Note on Plotly's Interactive Features

Plotly generates interactive plots that allow users to zoom, pan, and hover over data points for more information. This makes it particularly useful for exploratory data analysis and creating dashboards for stakeholders to interact with the data.

Note on Plotly's Subplot Specifications

When creating a dashboard with mixed plot types in Plotly, it's important to specify the correct subplot type for each plot. In this example, we use 'domain' for the pie chart and 'xy' for the other plots. This ensures that each plot is rendered correctly within its designated area in the dashboard.