Data and codes for the publication: Decay radius of climate decision for solar panels in the city of Fresno, USA

This repository contains core codes and data underlying the analyses and figures from: Barton-Henry, Wenz, and Levermann (2021).The following provides a brief description of the codes and data included.

Data:

fresno_merged_solarpanel_socioeconomic_dataset.csv - This file contains the core dataset used for the all results and figures. This dataset is a result of the aggregation of the following publicly available datasets:

• Bradbury, K. et al. Full collection: Distributed solar photovoltaic array location and extent data set for remote sensing object identification. Sci. Data https://doi.org/10.6084/m9.figshare.c.3255643.v2 (2016).
• The County of Fresno, California. GIS Shapefiles. Fresno County GIS Portal (2020) Retrieved from: https://www.co.fresno.ca.us/ departments/public-works-planning/divisions-of-public-works-and-planning/cds/gis-shapefiles, accessed January 2020.
• Institute of Education Science National Center for Education Statistics Education Demographic and Geographic Estimates (EDGE). School District Boundaries (2020). Retrieved from: https://nces.ed.gov/programs/edge/Geographic/DistrictBoundaries, accessed March 2020.
• U.S. Census Bureau. 2009–2013 5-Year American Community Survey Demographic and Housing Estimates (2014). Retrieved from https://data.census.gov/cedsci/table?q=acs&tid=ACSDP5Y2013.DP05&hidePreview=false, accessed November 2020.

variable_description.csv - This file includes a short description of each variable present in the main dataset, found in the fresno_merged_solarpanel_socioeconomic_dataset.csv file. 

Code included, in relevant order:

figure_1.py - This script produces Figure 1, a map showing the geographic area of analysis along with the geolocations of solar panels, addresses, as well as examples of several radii over which panel density is calculated.

figure_2.py - This script produces Figure 2, as well as outputs several dataframes with the normalized panel density variables used in the further analysis of the feature importances scores.  

figure_3.py - Figure 3 is produced, as well as several .csvs containing dataframes with varying panel density radii, constructed with the normalized panel density at one radius subtracted from that of the previous.

figure_4.py - Produces Figure 4, an analysis of the influence of household income. Figures S15 and S17 of the Supplementary Information section are also created. 

models.py - This script builds the three different models tested (Random Forest, AdaBoost, and XGBoost), and calculates the feature importances scores as well as performance metrics for each. 

confusion_matrices.py - This script creates figures of the confusion matrices created from the performance output for each of the tested models, as well as the model for which panel density is omitted. These are Figures S2, S4, S6, and S7. 

ols.py - This script produces the estimates contained in Tables S7, S8, and S9, which are the results of the OLS analysis.

tract_ave_plot.py - Figures S8 and S9 are produced, showing changes in the feature importances when panel density is averaged over census tract. 

households_all.py - Analysis of the influence of density based on the number of households in each tract, resulting in Figures S11 and S12. 

tract_area_all.py - Creates Figures S13 and S14, providing the analysis of the influence of density across census tracts of varying area sizes. 

hhval_all.py - The analysis of the influence of household value is conducted, and Figure S16 is produced.