Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Developing Cost Functions for Estimating Solar Photovoltaic System Installed Using Historical Data and OLS Regression



David H. Webb, Joshua D. Kneifel


Solar photovoltaics (PV) continues to increase in market share. Policy decisions and the nature of solar markets continue to shift; however, it is likely that the price of solar will continue to decrease in the near term. Given the increasing market and more competition in installations, it is beneficial to have a greater understanding in the driving factors in solar PV pricing, as well as models to help perspective buyers and sellers to obtain estimates for the cost of installations. Currently, most estimates rely on a marginal cost that is equivalent to the total cost divided by the system size. This study uses data from EnergySage and the National Renewable Energy Laboratory’s Tracking the Sun data set for California, specifically Fresno, San Francisco, Los Angeles, San Diego, and San Jose, to accomplish three goals: to determine if there are significant predictors for solar PV pricing outside of the current method of relying on system size only, to determine what model would make sense for predictive purpose in preparation for the development of a tool to predict the total life cycle cost of solar PV, and to determine if smaller geographical resolutions are warranted when looking at price by location. This paper finds that there are several more significant predictors of Solar PV pricing by including more PV system specifications, such as panel efficiency, inverter type, and system quality. Results also indicate that the installer of the PV system may proxy for the specification variables when it is included in the model. While the installer-based models show significant difference from many of the other models, including the specification-based models, they fail to increase the predictive capability for the EnergySage data, however, show promise for better predictions using the Tracking the Sun data. By breaking the data down to models by city and city-installer groups regional differences can be clearly seem, indicating a more refined geographic approach is ne
Technical Note (NIST TN) - 2114
Report Number


Cost Estimation, Economics, Regression Analysis, Solar Photovoltaics


Webb, D. and Kneifel, J. (2020), Developing Cost Functions for Estimating Solar Photovoltaic System Installed Using Historical Data and OLS Regression, Technical Note (NIST TN), National Institute of Standards and Technology, Gaithersburg, MD, [online], (Accessed May 24, 2024)


If you have any questions about this publication or are having problems accessing it, please contact

Created November 19, 2020, Updated November 20, 2020