Developing Data Products Project - MPG Predictor Application

This application can be used to predict the miles per gallon given the weight and horsepower of a vehicle.

The application is simple:

• Enter the weight, in half-tons (the default value is the mean of the mtcars dataset)
• Enter the horsepower (the default value is the mean of the mtcars dataset)
• Click “submit” and view the results
• Results are displayed graphically on two plots for comparison to mtcars values

The calculations were built using the mtcars dataset, so the application is intended for use on older vehicles.

The MPG is predicted from a linear regression model calculated from the mtcars dataset.

fit <- lm(mpg ~ wt + hp, data = mtcars) #regression model

            Estimate Std. Error t value  Pr(>|t|)
(Intercept) 37.22727    1.59879  23.285 2.565e-20
wt          -3.87783    0.63273  -6.129 1.120e-06
hp          -0.03177    0.00903  -3.519 1.451e-03

The previous code results in the following formula:

Let
\( \beta_0 \) = an intercept value, 32.22727
\( \beta_{wt} \) = the weight coefficient, -3.878
\( \beta_{hp} \) = the horsepower coefficient, -0.0318
\( X_{wt} \) = weight, in half-tons
\( X_{hp} \) = horsepower

Then

\( MPG = \beta_0 + \beta_{wt}X_{wt} + \beta_{hp}X_{hp} \)

Where MPG is the predicted miles per gallon