Use cases

COVID-19 quarantines and domestic violence in Bogota

COVID-19 quarantines have had multiple social effects that have a potential increase in domestic violence cases.

We use quarantine data and a quantification of mobility changes to develop a spatio-temporal model to understand the association of mobility patterns with incidents of domestic violence in Bogotá, Colombia. In addition, we estimated different regression models to study the causal relationship between quarantine duration and domestic violence. This project was conducted by researchers from Quantil and the Secretariat of Security, Coexistence and Justice of Bogota with support from the United Nations Development Program for Latin America and the Caribbean and GRANDATA.

Approach

We used information on imposed mobility restrictions and quantified mobility changes to develop a spatio-temporal domestic violence occurrence prediction model coupled with machine learning interpretation techniques to understand the association of mobility patterns with the spatio-temporal behavior of domestic violence incidents in Bogotá. In addition, we estimated different OLS-based regression models to better understand the causal relationship between the duration of quarantine blocking and domestic violence, which we initially found to be significant from the spatio-temporal predictive approach.

We work with daily and spatially aggregated observations of domestic violence from the Statistical, Criminal, Contraventional and Operational Information System of the National Police (SIEDCO) and the Purple Line of the District. These observations are aggregated into urban sectors of approximately 0.5 km2 , with 623 spatial divisions over the urban region of Bogotá. For each sector, we calculate whether there was at least one incident reported by SIEDCO or the District's Purple Line. Additionally, we used mobility data provided by Grandata to quantify the degree of compliance with quarantine restrictions in Bogota along with mobility restrictions imposed in Bogota between March 19 and September 9, 2020.

Results

Both the GAM and OLS models found a positive relationship between quarantine confinement and the reported incidence of domestic violence. Specifically, we found that:

  • Longer confinements are causally related to a higher incidence of domestic violence reports.
  • The relationship between quarantines and reports of domestic violence is not primarily driven by changes in mobility per se.
  • The increase in reports of domestic violence is probably due to other factors that were affected by the confinement, such as psychological and economic factors.
  • The OLS model of heterogeneous effects showed that the effects were driven by the two lowest social classes.

Other projects

In Projects, you can learn about real applications of our services, use cases and examples of them.

Pensiones1

Contractual Design for the Management of FONPET

Energia_electrica

Systemic Risk of Electricity Supply

planes-desarrollo1

Inventory Logistics - Multimodal Transportation

transporte_multimodal1

Integrated Development Plans

estrategia_de_deuda_banner

Ministry of Finance and Public Credit (MHCP) defines a Debt Management Strategy

reportes

Asset Allocation Model in an insurance company