Tools for Analyzing NIBRS Data and Tips for Merging NIBRS Data with Publicly Available Datasets Scott Came Executive Director Becki Goggins Director, Law and Policy Program 2016 Association of State Uniform Crime Reporting Programs (ASUCRP) Conference September 29, 2016
SEARCH, The National Consortium for Justice Information and Statistics Nonprofit organization established in 1969 Governed by a Membership Group of governor appointees from the 50 States, the District of Columbia, and the territories Policy statements are available at: http://www.search.org/about-search/policystatements/
SEARCH s NCS-X Work Technical readiness assessments Required for Bureau of Justice Statistics funding Train-the-trainer model NIBRS Technical Assistance NIBRS Toolkit
Tableau Software Tableau Public is available as a free download Installations available for Mac or PC Visualizations (or vizzes ) are hosted on cloud at https://public.tableau.com Users can email links to galleries to others to view Full version is available at no cost to students and teachers
Seattle Dashboard
Seattle Maps
2015 Seattle Offenses by Month
Burglaries by Month
Bicycle Thefts by Month
Burglaries by Day of Week
Bicycle Thefts by Day of Week
Bicycle Thefts v. Other Thefts By Day of Week
Burglaries by Time of Day
Bicycle Thefts by Time of Day
Assaults and Disturbance Offenses By Time of Day
Combining NIBRS with Other Data Vast Amounts of Open Data Freely Available Tools Data Science Skills Greater value from NIBRS!
Case Study 1: Crime and Weather Current Directions in Psychological Science (2001) Hot temperatures increase aggression by directly increasing feelings of hostility and indirectly increasing aggressive thoughts. 2 degrees Fahrenheit (Avg Annual Temp) 9 murders+assaults per 100,000 population but what can we learn about this from NIBRS data?
Case Study 1: Approach Our Goal: Heat Hypothesis Test Crime Rate Crime Rate Data Temperature Data Get lots of observations Join them together Visualization Regression analysis Temperature
A Brief Aside: Tools Working with diverse, often messy, datasets requires powerful tools and skills You can only go so far with Excel If is not on your radar it should be (whether you are an IT person or not) Open Source Freely Available Awesome graphics/charting GIS Stats O Reilly Data Science Survey (2016)
Case Study 1: Crime Data NIBRS provides us with: Daily crime counts* Rich (fine-grained) classification of crimes Agency identifier (can be roughly mapped to county) Use your own curated NIBRS data ICPSR https://www.icpsr.umich.edu/icpsrweb/icpsr/series/128 *When actually entered daily by the LEAs
Case Study 1: Crime Data Generally good to start with an initial visualization: Annual NIBRS offenses per 10,000 residents, 2013-2014
Case Study 1: Temperature Data Global Historical Climatology Network GHCN Temperature Precipitation Weather Conditions More 54,000+ weather stations globally Excellent US coverage back to the early 1900s Daily observations Station file gives lat/long of each stations
Case Study 1: Join Data Aggregate NIBRS data by day and county (we chose urban counties in the two states) Select GHCN stations with regular daily obs Find GHCN station closest to the centroid of each county (US Census provides centroids) Data quality: furthest station was 51 miles Aggregate temperature data for each county s station Merge NIBRS and weather data together
Case Study 1: Visualization
Case Study 1: Regression Ohio Washington
Case Study 2: Bars and Crime Question: Is a higher density of bars/taverns in an area correlated with crime rate? Not much research on this Survey paper by Pacific Institute for Research and Evaluation http://resources.prev.org/documents/alcoholviolencegruenewald.pdf
Case Study 2: Data NIBRS data again, aggregated by county (2013-2014) What about bar/tavern density??
Case Study 2: OpenStreetMap Crowd-sourced GIS repository Streets Geo-political boundaries Places https://www.openstreetmap.org Amenities like restaurants, parks, pharmacies, gas stations and bars/taverns Provides an application programming interface (API) to expose the data for free
Case Study 2: Visualization
Case Study 2: Interpretation Hypothesis: OpenStreetMap amenity data are incomplete and degree of completeness varies across areas
Final Tips Analyzing NIBRS data and merging datasets does not have to be expensive Many toolsets are readily available at no cost Require various skill levels Various levels of functionality Consider opportunities for partnering Statistical Analysis Centers Universities
NIBRS Toolkit Open-source analytics tool Open-source NIBRS validation tool Configurable for any state or jurisdiction Enforces all FBI NIBRS edits and validation rules Designed to speed transition to NIBRS
Thank you! Scott Came scott@search.org Becki Goggins becki@search.org For more information about SEARCH, please visit: www.search.org. SEARCH, The National Consortium for Justice Information and Statistics search.org