Correlation Heatmaps

Part of the Statistics for ML section of Coddy's Machine Learning journey. Lesson 32 of 64.

A matrix of twenty features is a wall of numbers. A heatmap colors each cell, so the strong pairs jump out. A cinema's eight days:

A 4 by 4 heatmap titled Cinema Correlations with a color bar from -1 to 1. Tickets and popcorn are deep red together, temp_c is deep blue against both, and the staff row is pale

This code draws it:

cinema = pd.DataFrame({
    "tickets": [120, 340, 210, 410, 180, 290, 380, 150],
    "popcorn": [70, 150, 100, 200, 95, 120, 170, 60],
    "temp_c": [24, 12, 15, 9, 20, 19, 11, 22],
    "staff": [6, 5, 8, 7, 5, 6, 6, 8],
})
corr = cinema.corr()
fig, ax = plt.subplots()
im = ax.imshow(corr, cmap="coolwarm", vmin=-1, vmax=1)
fig.colorbar(im)
ax.set_xticks(range(4), corr.columns)
ax.set_yticks(range(4), corr.columns)
ax.set_title("Cinema Correlations")
plt.show()

vmin=-1, vmax=1 pin the colors: red is 1, white 0, blue -1. fig.colorbar(im) adds the key, and set_xticks(positions, labels) names the cells.

Tickets and popcorn glow red together; warm days are blue against both. Staffing barely matters.

Challenge

Medium

A smartwatch logged 30 days in watch.csv: a text weekday column and five number columns.

Compute the correlation matrix of the number columns.
Draw it as a coolwarm heatmap from -1 to 1, with a color bar.
Label the ticks across and down with the column names, title it Watch Data, and show it.

Your program prints nothing. For another owner's three columns, the chart looks like this:

A 3 by 3 heatmap titled Watch Data with steps, sleep_h and stress as tick labels and a color bar from -1 to 1: red diagonal, sleep_h and stress deep blue, steps and sleep_h light red

Try it yourself

import pandas as pd
import matplotlib.pyplot as plt

watch = pd.read_csv("watch.csv")

# Write your code below

Test yourself

This lesson includes a short quiz. Start the lesson to answer it and track your progress.

All lessons in Statistics for ML

3Probability by Simulation

Probability as Frequency The Law of Large Numbers Sums of Dice Expected Value Conditional Probability Recap: Dice Stall Odds

6Relationships

Reading Scatter Plots Correlation Correlation MatrixCorrelation HeatmapsCorrelation Is Not Cause Recap: Ice Cream and Sun

9Feature Scaling

Why Scale Features Min-Max Scaling Standardization StandardScaler Recap: Climbing Partners

12Final Challenges

Train Delays Air Quality Wine Chemistry Clinic Survey

Practice on your own: Online Python compiler