This repository includes the code for the analysis implemented in the analysis of the Living with Floods data, distributed by the World Bank.
The data and code in this repository include the raw data in CSV format, data preparation and cleanup scripts in Python, Stata DO file with logit regression model scripts, and the resulting models.
Additional data includes spreadsheets in Excel file format with the households and individual scale databases.
This analysis uses logit models to evaluate the effect of risk perception and response capacity during a flood event in Porto Alegre, Brazil, at the margins of the Jacuí river delta.
We developed three sets of regression models with risk perception as dependent variable and step-wise added effects of gender, number of residents in household, monthly household income, and age. We tested the effects of the dependent variable and co-variates against 14 risk response options that the residents of the affected areas took during the flood. These responses include Evacuate (RP1), Turn off power (RP2), Place obstacles to block water entry (RP3), Adjust interior (RP4), Adjust exterior (RP5), Stock food reserves (RP6), Seek information (RP7), Join a community alert group (RP8), Ask for help (RP9), Provide information (RP10), Pray (RP11), Other (RP12), Nothing. there was no time (RP13), Prepare before the flood (RP14).
The results of the analysis are presented and discussed in a forthcoming publication.