Econometrics Report: Multiple Linear Regression of Paris House Prices

This report applies a multiple linear regression model to analyze the factors influencing house prices in Paris. The analysis utilizes a dataset comprising 54 observations across nine districts, incorporating variables such as the number of rooms, area in square meters, district location, luminosity, renovation status, ground floor presence, garden/terrace availability, parking, safety, isolation, access to transportation, and proximity to bars/restaurants. The study begins with data preparation, including coding categorical variables, followed by descriptive statistics for scale variables, and a forward selection procedure to identify significant variables. The regression model, which excludes the constant term, is then presented, and the model's fit is evaluated using R-squared and F-statistic. The report then examines the individual variable coefficients, assessing their significance and collinearity using VIF values. The analysis was conducted using Minitab due to the mixed variable types. The multiple regression equation is then presented, and the model's assumptions of linearity, independence, normality, and equal variance are validated through residual analysis. The findings suggest that the number of rooms, area, and district location significantly influence house prices in Paris. Finally, the report concludes that the model is appropriate for estimating house prices in Paris, and that the high demand for houses in Paris might make other variables less significant in the pricing of the house.