Models for evaluating the impact of interventions on time series

• Grant number: 25/05464-2
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: June 01, 2025
• End date: May 31, 2028
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Airlane Pereira Alencar
• Grantee: Fabio Jose Guedes Magrani
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated research grant: 23/12851-7 - Evictions Observatory: interdisciplinary dimensions of risk in times of climate change and housing crisis, AP.TEM

Abstract

The aim of our project is to evaluate the impact of public policies on climate variables, socio-economic indicators and, if possible, indicators constructed from data collected from social networks and internet citations that can capture factors related to vulnerability, especially in the area of housing. In general, time series have various components, such as trend, seasonality, effects of explanatory variables, and even changes in trend and different effects. Any proposal for statistical models must take into account the nature of the data, which can be a count, a continuous non-negative variable, and does not necessarily have to have a normal distribution. In particular, for time series, we must consider that the observations measured at a given point in time are correlated with previous observations, which we call autocorrelation. Various models based on SARIMA (Box and Jenkins, 1970), GARMA (Benjamin et al., 2003), GAMAR (Yang et al., 2012) will be adjusted depending on the behavior of the series analyzed.