For energy star rated building, the adjusted R^2 for 2013 data is 0.978 which means the linear fit is passing through majority of the data and is a good fit. Same conclusion can be made for the 2014 data with adjusted R^2 equals 0.975. In the appendix below both both OLS plots show data below the median, 50 or less stars and above the median 50 or more energy data. The blue line represents the below median and the green line above. Although both data seem to have good statistical distribution and fit, it seems they look dissimilar.This may due to outlier points embedded in the data source. The point-log6 on the x-axis on the 2014 plot is an example. Additionally, the 2014 data has much more rows, about 1500 to 7000 rows, than 2013 data set. This may introduce more buildings per zip-code area than in 2013 and large differences of EUI among buildings. Also, the comparison of line fits in 2013 yields that green line intersects EUI axis approximately at log3.8 whereas blue passes through at about log5.5. This may indicate that buildings with 50 or more stars have less impact on energy resources than those with less stars. However,, 2014 data is difficult to interpret due to the fact that the blue line crosses the x-axis at zero and makes it unreliable to make predictions. With 2014, it would need a further study and removal of outliers from the data and apply OLS again.