Hw Ames Submission
2025-04-17
Introduction
The goal of my regression analysis is to find variables that help predict sales price of homes. I believed that lot area, gr_liv_area, and year_built would be good indications for sales price. I ended up with an R squared value of 0.664 and an RSME value of 46112, meaning on average the variation of a prediction is off by $46,112.
Data Cleaning and Minipulation
I noticed that the sales price column was skewed in one direction which would throw off my correlations. I fixed this by putting it on a Log scale which gave it a normal distribution. I also tried turning the street column into flag values to see a correlation but decided to not include it as a variable.
Variables
- lot area: This variable shows how large the property size is
- gr_liv_area: This variable gives the square feet of above ground space in the house
- year_built: This variable gives the year the house was built
Correlation and Outputs
The correlation between these variables and sales price isn’t the best but it did have some sort of shape. Below is a visual showing the shpe of the relationships. There is also a correlation matrix showing the values of how correlated they are.
