On Demand Presentations

Spatio-temporal modelling of COVID-19 data from the Indian State of Kerala

Presented by: Divya Kappara, University of Hyderabad

We propose spatial models to explain the district-wise COVID-19 incidence in a given region, using the airports as spatial hubs and the monthly air passenger traffic data at these airports. These models combine over-dispersed count data models along with ingredients of a gravity model with a suitable distance decay function, combined together in a generalized linear model framework.

For illustration purposes we have chosen COVID-19 incidence data from the Indian state of Kerala. Kerala was the first affected state and still continues to be one of the most affected regions in India. We have applied our model to the district level monthly COVID-19 data for the 14 districts and the monthly air passenger traffic data of the 4 international airports situated in Kerala since April, 2020. Our results suggest a significant effect of the air passenger traffic on the Kerala district level COVID-19 incidence.

View presentation - https://youtu.be/XTQqv646GcE

Modeling Occupancy and Detection of Indy Urban Wildlife

Presented by: Rasitha Jayaseare, Butler University, Indianapolis, IN

“Indy Wildlife Watch” is a project that monitors the wildlife species living in urban and peri-urban greenspaces within and just outside of the city of Indianapolis. The project deploys motion-triggered cameras at city parks, forest remnants, riparian forests, golf courses, agricultural lands, schools, and cemeteries spanning an urban-to-peri-urban gradient from downtown Indianapolis through the northern suburbs. Data on species distribution and their habitat use are extracted from the pictures taken during four months representing each of four seasons annually at about 50 sites. The occupancy and detection of each species are modeled using a series of Zero-Inflated regression models with covariates distance to city center and habitat type. This poster presents the model, results, and the impact of distance to city center and habitat type on the predicted occupancy and detection of urban wildlife in Indianapolis. Through further analysis, covariate information can be utilized to predict occupancy of a species given a particular environmental attribute and utilized to inform urban land management practices.

View presentation - https://youtu.be/XGvHlSeyJhk