The Covid-19 Seattle Street View Campaign
Abstract
In January 2020, Covid-19 was first detected in the United States outside of Seattle. By March 2020, Seattle and Washington State responded to the emerging pandemic with a series of restrictive public health measures, including limiting the size of gatherings and issuing the first stay-at-home orders. In response, a team of multidisciplinary University of Washington investigators, in partnership with the U.S. National Science Foundation (NSF)-supported Natural Hazard and Disaster Reconnaissance (RAPID) Facility, designed and implemented a longitudinal (time-series) "street view" campaign to collect image data tracking the pandemic's impacts and the community's recovery. In May 2020, the Covid-19 Seattle Street View Campaign was launched to produce an unprecedented, high-resolution, open, ground-based record of the urban region during and after a significant global health crisis. The campaign has collected imagery data approximately every three to six weeks along a designated 100-mile route through the city. In addition to collecting data to support the team's research interests and objectives, we explicitly sought to advance the scientific application of post-event mobile imaging by establishing sampling protocols that may be used to guide campaigns for future disruptive events. As the effects of the pandemic carry on, we anticipate the campaign will continue through at least April 2023. In this presentation, we describe the sample in the strategy employed in the research and discuss artificial intelligence-based schemes to interpret this vast urban time series imagery data set, which is openly available to the research and practice communities on designsafe-ci.org and streaming on the street image viewing platform Mapillary. We also present a series of initial findings from our street view data analysis to describe the effects of the pandemic and policy measures across a broad range of communities, services, and institutions in Seattle.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFMGH53A..09W