Spearman correlation in SciPy - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is Spearman correlation?

Spearman correlation measures how well the relationship between two variables can be described using a monotonic function. It uses ranks instead of actual values.

beginner

How does Spearman correlation differ from Pearson correlation?

Spearman correlation uses ranks and measures monotonic relationships, while Pearson correlation uses actual values and measures linear relationships.

beginner

Which Python library and function can you use to calculate Spearman correlation?

You can use the scipy library and the function scipy.stats.spearmanr to calculate Spearman correlation.

intermediate

What does the output of scipy.stats.spearmanr include?

It returns two values: the Spearman correlation coefficient (a number between -1 and 1) and the p-value that tests if the correlation is significant.

intermediate

Why is Spearman correlation useful with non-linear data?

Because it uses ranks, Spearman correlation can detect monotonic relationships even if they are not straight lines, unlike Pearson correlation.

What type of relationship does Spearman correlation measure?

A Monotonic relationship

B Only linear relationship

C No relationship

D Exponential relationship

Spearman correlation measures monotonic relationships, which means variables move in the same direction but not necessarily at a constant rate.

Which function calculates Spearman correlation in scipy?

A scipy.stats.spearmanr

B scipy.stats.kendalltau

C scipy.stats.pearsonr

D scipy.stats.linregress

The function scipy.stats.spearmanr calculates Spearman correlation.

What does a Spearman correlation coefficient of 1 mean?

A Perfect negative monotonic relationship

B Perfect positive monotonic relationship

C Perfect linear relationship only

D No relationship

A coefficient of 1 means a perfect positive monotonic relationship between variables.

Why might you choose Spearman over Pearson correlation?

A You want to fit a regression line

B Data is normally distributed

C You want to measure variance

D Data has outliers or is not linear

Spearman is better when data is not linear or has outliers because it uses ranks.

What does the p-value from scipy.stats.spearmanr indicate?

A Strength of correlation

B Probability correlation is zero

C Significance of correlation

D Direction of correlation

The p-value tests if the observed correlation is statistically significant or could happen by chance.

Explain how to calculate and interpret Spearman correlation using scipy.

Think about ranking data and what the two outputs mean.

Describe when and why you would use Spearman correlation instead of Pearson correlation.

Consider the shape of data and sensitivity to outliers.