Dylan Fitzpatrick

Dylan J. Fitzpatrick, Wilpen L. Gorr, Daniel B. Neill

Hot-spot-based policing programs aim to deter crime through increased proactive patrols at high-crime locations. While most hot spot programs target easily identified chronic hot spots, we introduce models for predicting temporary hot spots to address effectiveness and equity objectives for crime prevention, and present findings from a crossover experiment evaluating application of hot spot predictions to prevent serious violent crime in Pittsburgh, PA. Over a 12-month experimental period, the Pittsburgh Bureau of Police assigned uniformed patrol officers to weekly predicted chronic and temporary hot spots of serious violent crimes comprising 0.5 percent of the city's area. We find statistically and practically significant reductions in serious violent crime counts within treatment hot spots as compared to control hot spots, with an overall reduction of 25.3 percent in the FBI-classified Part 1 Violent (P1V) crimes of homicide, rape, robbery, and aggravated assault, and a 39.7 percent reduction of African-American and other non-white victims of P1V crimes. We find that temporary hot spots increase spatial dispersion of patrols and have a greater percentage reduction in P1V crimes than chronic hot spots but fewer total number of crimes prevented. Only foot patrols, not car patrols, had statistically significant crime reductions in hot spots. We find no evidence of crime displacement; instead, we find weakly statistically significant spillover of crime prevention benefits to adjacent areas. In addition, we find no evidence that the community-oriented hot spot patrols produced over-policing arrests of minority or other populations.

Chelsea Sidrane, Dylan J Fitzpatrick, Andrew Annex et al.

Flooding is a destructive and dangerous hazard and climate change appears to be increasing the frequency of catastrophic flooding events around the world. Physics-based flood models are costly to calibrate and are rarely generalizable across different river basins, as model outputs are sensitive to site-specific parameters and human-regulated infrastructure. In contrast, statistical models implicitly account for such factors through the data on which they are trained. Such models trained primarily from remotely-sensed Earth observation data could reduce the need for extensive in-situ measurements. In this work, we develop generalizable, multi-basin models of river flooding susceptibility using geographically-distributed data from the USGS stream gauge network. Machine learning models are trained in a supervised framework to predict two measures of flood susceptibility from a mix of river basin attributes, impervious surface cover information derived from satellite imagery, and historical records of rainfall and stream height. We report prediction performance of multiple models using precision-recall curves, and compare with performance of naive baselines. This work on multi-basin flood prediction represents a step in the direction of making flood prediction accessible to all at-risk communities.

Bhuwan Dhingra, Zhong Zhou, Dylan Fitzpatrick et al.

Text from social media provides a set of challenges that can cause traditional NLP approaches to fail. Informal language, spelling errors, abbreviations, and special characters are all commonplace in these posts, leading to a prohibitively large vocabulary size for word-level approaches. We propose a character composition model, tweet2vec, which finds vector-space representations of whole tweets by learning complex, non-local dependencies in character sequences. The proposed model outperforms a word-level baseline at predicting user-annotated hashtags associated with the posts, doing significantly better when the input contains many out-of-vocabulary words or unusual character sequences. Our tweet2vec encoder is publicly available.