Emergency services play an essential part in serving our communities and keeping them safe. The management of emergencies is a collaborative effort amongst various stakeholders. Real-time data provides new opportunities for improving this collaboration, b

In this talk, Dr. Delmelle will give three examples of how text and natural language processing methods can serve to complement traditional forms of quantitative spatial data and provide insight onto urban processes. All three examples use the public remarks that accompany real estate property listings, or the text descriptions of the properties. The first combines more traditional, quality of life data collected by the city of Charlotte, North Carolina with real estate text to explore neighborhood and housing characteristics of places that attract a relatively larger share of recent in-movers from outside of the MSA as compared to local movers. The second explores language associated with transportation amenities extracted from the advertisements, over time, in tandem with changes in the racial and income profile of mortgage applicants to a neighborhood. This highlights how terms associated with more sprawling forms of development (car, cul-de-sac) have been increasingly associated with more minority homebuyers while 'smart growth' associated terms (transit, walkability) have coincided with a dramatic uptick in whiter and wealthier applicants. Finally, the third example uses the text to classify listings on a continuum of housing investment and disinvestment to study housing dynamics at a small spatial and temporal resolution.