# PUI2016 Extra Credit

Text
Ozgur L. Akkas
PUI 2016
Extra Credit Assignment
12/15/2016

Problem Description:

Does Leadership in energy star buildings in NYC show substantial energy savings compared to those non-energy star buildings? Looking at the 2013 and 2014 NYC energy and water disclosure benchmark data, the City’s building can be evaluated based on the grade level they received. Energy start buildings can help reduce energy consumption and reduce environmental impact once a building meets Environmental Protection Agency regulations. According to the Department of Energy, to earn a certification, a building must earn Energy Star score of 75 or higher whereas a score 50 or above reflect the median energy performance[1]. To see whether energy star buildings units, this report compared buildings with the below median and those above the median.
Data:
New York City Benchmarking provides water and energy use data as CSV files collected during the year 2013 and 2014. In the analysis, weather normalized data is chosen as the dependent variable and the reported building floor area as the independent variable. Then, NYC zip codes data given in the same tables were used to stack the information to make statistical tests. Links to data are given in the Appendix.
Analysis:
This analysis uses ARIMA method for evaluation of the data sets. To find out the variance, adjusted R^2 value is calculated. R value was calculated to see whether the median line passes as many data points or not. When R equals zero, it means not the fit does not pass as many points and when R equals one line passes as many point. The Chi square is also calculated to evaluate the data error estimate to acceptable or too good. Chi of one means error estimate is acceptable and zero means error estimate is too unrealistic.
Results
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. Same conclusion can be made for the 2014 data with adjusted R^2 equals 0.975. For buildings with greater than 50 stars, the R is 0.980 for 2013 and 0.978 for 2014, respectively. In the appendix below both OLS plots show data below the median and above the median energy data for two consecutive years.
Recommendations for Further Analysis:
Clustering of the data with the energy use intensity (EUI) and apply a geospatial distance map to see how energy star certification has impacted various City neighborhoods. Also, producing a time series analysis beginning from a certain year when energy star certification for building has been issued can be another area to be studied to measure the impact on buildings’ total green house gas emissions and the future trend. ARIMA can be used to make predictions and estimate.

Appendix:
Benchmarking Data CSV files