DataCamp

Graphically evaluate the unemployment model

In this exercise you will graphically evaluate the unemployment model, unemployment_model, that you fit to the unemployment data in the previous chapter. Recall that the model predicts female_unemployment from male_unemployment.

You will plot the model's predictions against the actual female_unemployment; recall the command is of the form

ggplot(dframe, aes(x = pred, y = outcome)) + 
       geom_point() +  
       geom_abline()

Then you will calculate the residuals:

residuals <- actual outcome - predicted outcome

and plot predictions against residuals. The residual graph will take a slightly different form: you compare the residuals to the horizonal line \(x=0\) (using geom_hline()) rather than to the line \(x=y\). The command will be provided.

The data frame unemployment and model unemployment_model are available in the workspace.

Use predict() to get the model predictions and add them to unemployment as the column predictions.
Plot predictions (on the x-axis) versus actual female unemployment rates. Are the predictions near the \(x=y\) line?
Calculate the residuals between the predictions and actual unemployment rates. Add these residuals to unemployment as the column residuals.
Fill in the blanks to plot predictions (on the x-axis) versus residuals (on the y-axis). This gives you a different view of the model's predictions as compared to ground truth.

Take Hint (-30 XP)

Exercise

Graphically evaluate the unemployment model

Instructions