Predicting car prices
The purpose of the dissertation was to determine which factors affect the price of the car.
As soon as we collected the data, we presented it graphically. For example, in the graph below you can see the percentage of each fuel type in the data.
A linear regression model was used to identify which factors affect the price. This model can be expressed mathematically as:
where yi represents the dependent variable, xi independent variables, and u is error term. Regression parameters (also called Betas) were estimated by ordinary least squares (OLS) method.
The table below shows the estimated parameters:
The p-value of Mileage is lower than 0.05. It means that this factor is statistically significant (at 5% significance level) and it does affect the price of the car.
The negative linear relationship between price and mileage is visualised in the scatterplot below:
The higher the mileage, the lower the price.