Generate two coloured Scatter Plots: first mean radius versus mean perimeter, then mean radius versus mean area. Put radius on the x-axis in both plots. Inspect the shape before using any correlation coefficients.

Which observation is supported?

Now, for the mean area histogram, switch on Color by Target, so that it shows the distribution of mean area for benign and malignant samples.

Use the plot to answer the following questions. Approximate values are sufficient.

1. Typical values: In which range are most benign samples concentrated? How does this compare with the malignant samples? Which group typically has larger values?

2. Spread: Which group’s values are spread over a wider range? Support your answer by referring to the approximate ranges visible in the plot.

3. Shape and extreme values: Look at the malignant samples. Is the distribution roughly symmetric, or does it extend further towards smaller or larger values? A few samples have values above 2,000. Would you describe these as typical of the group? Explain why or why not.

4. Evaluate a statement: A student says: “Malignant samples always have a larger mean area than benign samples.” Do you agree? Look for ranges in which the two distributions overlap and use them to justify your answer. If necessary, rewrite the statement so that it better reflects what the histogram shows.

Attach the histogram here also.

Choose Pairwise Relationships with exactly these four features, in this order: mean radius, mean perimeter, mean area, mean texture. Keep Color by Target on. Read the axis labels and inspect the diagonal.

Which statement correctly describes this matrix?

Generate colored Violin Plots for mean radius and mean fractal dimension. View each feature separately if that makes its axis easier to read. For each feature, compare the positions and overlap of the two diagnosis groups.

Which description best matches these plots?

Create a scatterplot with mean radius on the x-axis and mean texture on the y-axis. First inspect it with Color by Target switched off.

What can you say about the following points?

1. Overall relationship: As mean radius increases, how does mean texture generally change? Do the points follow a narrow pattern, or is there considerable spread?

2. Diagnosis groups: Compare the locations of the Benign and Malignant samples. Which group tends to have larger values of the two features?

3. Overlap: Are the groups clearly separated, or do they overlap? Describe one region of the plot that supports your answer. Would the plot justify expecting an error-free diagnosis rule based on these two features alone?

Attach a screenshot of the colored scatterplot.

Keep mean area. Now switch Color by Target on. Look for the same Malignant sample with mean area = 1509. Where is it now? Do you have an idea why we see this effect?

Keep mean area. Create a Box Plot with Color by Target off. The dataset includes a Malignant sample with mean area = 1509. Search for it in the box plot. How do you interpret this specific sample with respect to how it is shown in the boxplot?