MechaCar Statistical Analysis

Linear Regression to Predict MPG

The vehicle length and the ground clearance as well as the intercept, are statistically unlikely to provide random amounts of variance to the linear model.

In addition, the linear model predicts MPG of MechaCar effectively because the p-value is less than 0.05 (5.35e-11) and the adjusted R-squared is close to 0.7 (0.6825)

Summary Statistics on Suspension Coils

The mean and the median are very close to each other, means the data set has a symmetrical distribution. The distribution chart also confirms this.

The variance and the standard deviation for the whole population are high. This means the data points are very spread out from the mean and from one another but meets the design specification for all manufacturing lots in total.

NOTE: The design specifications for the MechaCar suspension coils dictate that the variance of the suspension coils must not exceed 100 pounds per square inch.

All lots together meet the design specifications, with a variance of 62.29, but as an individual lots, only Lot1 and Lot2 meet the criteria, with variances of 0.98 and 7.47 respectively. Lot3 surpass the minimum requirement with a variance of 170.29, 70 PSI above the manufacturer specifications.

T-Tests on Suspension Coils

Using the T-Test we can conclude that the PSI across all manufacturing lots is 'Statistically Different' from the population mean of 1,500 pounds per square inch because the p-value is greater than 0.5 (0.6036).

Comparing each manufacturing lot against the mean PSI of the population, I can conclude that each lot is NOT statistically different from the population mean of 1,500 pounds per square inch because each lot has a p-value lower than 0.5.

Study Design: MechaCar vs Competition

Does a high horsepower over other manufacturer positively affect the sales of cars?

  • Hypothesis: Having a higher horsepower performance than other manufacturers will increase the sales of the model X.
  • Null Hypothesis: Having a higher horsepower performance than other manufacturers will show no improvements on the sales of the model X.

The data needed is the sales of hour model X and the other manufacturer cars in the same car range type before and after the increase of the model X horsepower. We can also analyze the changes of the horsepower over the past 10 years with similar cars and see if any change on the horse power affects the sales of our cars or the competition cars.