Analysis and Prediction of 911 Calls based on Location using Spark Big Data Platform
MetadataShow full item record
Type of Work16 pages
Citation of Original PublicationKetki Deshpande et al., Analysis and Prediction of 911 Calls based on Location using Spark Big Data Platform, http://hpcf-files.umbc.edu/research/papers/IS789_Project_Report-HPCF-2019-29.pdf
RightsThis item is likely protected under Title 17 of the U.S. Copyright Law. Unless on a Creative Commons license, for uses protected by Copyright Law, contact the copyright holder or the author.
Proper management of critical resources like Police Force and Ambulance Services is the key to establish peace quickly in times of crisis. When a police district receives a 911 call, quick response can be the difference between quiet handling and full riots in any area. Through this project, we have tried to determine frequent patterns for establishing association between day and time of the week, police district-based location and the reason for the call. We have also tried to predict the number of calls from a particular location (using longitude and latitude data). This data will help us manage police resources and put them to apt use as and when required. We have used Spark based algorithm known as FP-Growth for finding the frequent patterns in the calls and a couple of Regression algorithms Decision Tree and Random Forest for the prediction of calls based on location. Results show that weekend evenings are the busiest time for the emergency services, as most of the calls are made in the evenings of Friday, Saturday and Sunday. Also, Northwestern, Southwestern and Northeastern Police Districts get most of the calls in the evening. Based on the existing training data, we were able to predict new calls for a particular location.