Business Intelligence System for Traffic Data Integration: Linking Roadway, Collision and Traffic Flow Data to Improve Traffic Safety Stevanus A. Tjandra, Ph.D. City of Edmonton Office of Traffic Safety
Outline City of Edmonton Office of Traffic Safety Performance Measurement and Traffic Safety Problem Business Intelligence (BI) System for Traffic Data Integration Integrated Data Analysis BI, KPI and Predictive Analytics
City of Edmonton Office of Traffic Safety (OTS) http://www.edmonton.ca/transportation/traffic-safety.aspx Established in 2006 as a result of Mayor Traffic Safety Task Force World urbanization - UN report 2009: 1 Year Urban Population Rural Population 2009 3.4 billion 3.4 billion 2050 6.3 billion 2.9 billion The OTS will reduce the prevalence of fatal, injury, and property damage collisions through the 4 E s of traffic safety (Engineering, Education, Enforcement, and Evaluation) by improving urban traffic safety engineering road user behaviour speed management and data, business intelligence and analytics 1 World Urbanization Prospects The 2009 Revision http://esa.un.org/unpd/wup/documents/wup2009_highlights_final.pdf
Measuring Performances You Can t Manage What You Don t Measure What Gets Measured Gets Done (Peter Drucker) Data + Analytics Strategic Statement Key Result and Performance Indicators (KRIs and KPIs)
KPI Level I and II Level I Level II
Traffic Safety: A Multidimensional Problem Roadway Engineering Geometric design: curves, slope, lane width, etc. Non-reflective paint Speed bump Lighting Road physical conditions Enforcement camera: intersection, road segment Check Stop Traffic signs Presence of enforcement officers Regulation Speed limit Seat belt Road connection that creates high traffic variability Demand vs. capacity: road congestion, required # road connections to meet the traffic growth Brake malfunction Visibility of windows Road User Behaviour Alertness: alcohol/drug, fatigue, health Attention: cell phone, music Side-view mirrors Vehicle safety performance: crashworthiness, aggressivity. Distracted Driving Law Dynamic Message Sign (DMS) Speed display Age Gender Community Program to promote traffic safety Speeding Visibility Index Seat belt Animal crossing Driving decision Driving skill/training Presence of passengers Socioeconomic level Sun glare Ice, Rain, Fog Traffic safety Enforcement Vehicle Education Nature
Weather Impacts on Traffic Safety: Trends and Patterns
Repeat Violators
% Vehicles Involved in at Least 1 Collision Violations and Collisions 30% 25% 20% 15% 10% 5% 0% 1 2 3 4 5 6 7 8 9 10 11 12+ Number of AE violations Edmonton Automated Enforcement and Collision Data: January 1, 2010- December 31, 2011 Topinka, Neil. Project Mercury: Automated Enforcement Data and Driver Behaviour in the Edmonton Capital Region. International Forum on Traffic Records And Highway Safety Information Systems - Special Session, May 2, 2013. Edmonton, Canada.
Urban Traffic Safety Engineering Road Design Improvement Yellowhead Trail WB RAMP Victoria Trail NB
Before Urban Traffic Safety Engineering Video Analytics Yellowhead Trail WB RAMP Victoria Trail NB After The project cost: $437,240 The collision reduction resulted in an annual cost saving of $887,685
Data, Information, and Knowledge Silos Crime Inadequate knowledge about the existence of various data and their availability, Lack of linkages with other databases resulting in duplicate data collection, processing and management, No standardized method for the specific identification of attributes across data sources, Modified from the original picture published in http://blogs.sun.com/bblfish/entry/business_model_for_open_distributed Lack of communication among stakeholders of important changes to the data, and Lack of access to other data systems
Traffic Data Coordination Committee (TDCC) Policy direction, strategic needs Traffic Data Oversight Committee Program oversight Recommended Integrated System and Five-Year Plan Traffic Data Coordination Committee University of Alberta Policy direction, strategic and tactical needs Recommended operational needs and actions (e.g., standardized format, critical data, data collection and analysis procedures) Program oversight Task Force Task Force Task Force Task Force
Data Analysis and Business Intelligence Traffic Data Integration Automated Enforcement (Intersection Safety Cameras and Photo Radar Cameras) Traffic Count Management (TCM) Motor Vehicle Collision Information System (MVCIS) Enforcement durations Traffic counts Offence statistics Issued tickets statistics Locations Traffic volume Turning movement counts Speed surveys Roadway collisions Spatial Land Inventory Management (SLIM) Initial Release Roadway inventory Speed & traffic complaints Traffic Complaints System Spatial Business Intelligence Framework Road conditions Roadway maintenance Transportation Operations Weather conditions Demographic statistics Manned enforcement Impaired driving Crime statistics Shift schedules Accidents/incidents Bus Stops & routes Passenger counts Transit Environment Canada Federal Census Edmonton Police Service Transit Safety & Security 2011 WINNER of IBM SMARTER CITY, one of 24 cities worldwide
Spatial Business Intelligence Spatial Business Intelligence (SBI) supports traditional and spatial data through merging Geographic Information Systems (GIS) and Business Intelligence (BI) technologies Analytics Mapping Reporting EFFICIENT & EFFECTIVE DECISION MAKING Integration &Collaboration
Business Intelligence of Traffic Data Integration Use BI Launchpad in BusinessObjects 4.0 as an Enterprise Report Portal for sharing business intelligence reports
Business Intelligence of Traffic Data Integration The last traffic survey: 2011
Benefits of Business Intelligence (Eckerson, The Data Warehousing Institute, 2003)
Integrated Data Analysis: Inner Ring Road Empirical Bayes analysis led us to identify 170 St - North of 95 Ave as the worst section for traffic collisions From 2009 to 2011 there were: 55% more collisions than other similar mid-block locations 23 more collisions than expected West Edmonton Mall *Brandt Denham, City of Edmonton Office of Traffic Safety, 2013
Integrated Data Analysis: Inner Ring Road Collisions by Cause (09-11) Struck Parked Veh 2% Flwd. Too Closely 72% Ran Off Road 2% Chng. Lanes Impr. 18% Fld. Yield R.O.W. 6% Collisions by Driving Lane (2012) 25 20 15 10 5 0 Collisions by Day of Week (09-11) Mon Tue Wed Thur Fri Sat Sun Peak collision periods: Nov-Dec Christmas shopping Fri-Sat weekend shopping Mid afternoon shopping 1 2 3 Unknown 17% 2012 data has less unknown traffic lanes so it may be a more accurate breakdown of the collisions by lane Violator Registered Owner Postal Code 3rd from Curb 8% Right Curb 25% 2nd from Curb 50% The 2 nd from curb lane is lane #3 (The right curb lane is not a through lane) Within Edmonton 58.74% 41.26% Outside of Edmonton Average Monthly Speed Tickets Issued Average Monthly Red Light Tickets Issued Study area Lane 3 (67) 20% 23% Lane 1 (77) Lane 3 (10) 36% 24% Lane 1 (6) The top 5 violators are all rental and cab companies. Lane 2 (195) 57% 40% Lane 2 (11) Implementation
BI, KPI and Predictive Analytics
Predictive Analytics
Short-Term Weather and Collision Prediction
City of Edmonton Weather, Collision and Traffic Flow Prediction Weather Data Set-up for Edmonton Calendar Data Month, Day of Week, Holiday Traffic Flow Data VDS Sites: Speed, Volume Snow Fall Amount (cm) 16 14 12 10 8 6 4 1-day-earlier Forecast Snow (cm) Actual Amount Snow (cm) 7-day-earlier Forecast Snow (cm) Past Future 2 8 0 12-Feb 13-Feb 1-Apr 2-Apr 3-Apr 4-Apr 5-Apr 6-Apr 7-Apr 8-Apr 9-Apr Predicted Total & FTC Collisions Amount of Precipitation (MM) 90 85 80 75 70 65 60 Traffic Speed (km/h, Location: SW VDS Site - WMD & 156 St) High Speed Alert Warranted Future Work 11/20/2012 11/22/2012 11/24/2012 11/26/2012 11/28/2012 11/30/2012 12/02/2012 12/04/2012 12/06/2012 12/08/2012 12/10/2012 12/12/2012 12/14/2012 Volume & Speed Prediction 12/16/2012 Day 12/18/2012 12/20/2012 12/22/2012 12/24/2012 12/26/2012 12/28/2012 12/30/2012 01/01/2013 01/03/2013 01/05/2013 01/07/2013 01/09/2013 140 PredictedTotal PredictedFTC AmountPrecipitation 14-Feb 15-Feb 16-Feb 17-Feb 18-Feb 19-Feb 20-Feb 21-Feb 22-Feb 23-Feb 24-Feb 25-Feb 26-Feb 27-Feb 28-Feb 1-Mar 2-Mar 3-Mar 4-Mar 5-Mar 6-Mar 7-Mar 8-Mar 9-Mar 10-Mar 120 100 7 6 Weather 2 Prediction 0 80 60 40 5 4 3 0 20 1 Collision Prediction
The Growing Challenge Collisions in Edmonton 1997-2013 Total Collisions Injury and Fatal Collisions Collisions per 1,000 Population 34.8 35.2 38.5 38.6 36.8 35.9 28.2 30.1 29.4 31.9 31.5 31.7 29.1 32.0 28,520 29,072 28,832 28,480 28.9 28.4 29.7 20,992 21,000 23,542 22,137 20,606 22,784 26,066 23,442 23,243 24,803 19,128 19,082 17,648 5,665 5,951 6,327 6,817 7,151 7,658 6,381 5,564 5,873 6,092 Decrease 14,451 injury & fatal collisions (20,795 injuries and fatalities) Save $ 999.8 M 5,513 4,758 3,991 3,792 3,504 3,389 3,246 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Zero fatal and serious injury collisions Our Vision: Zero Fatal and Injury Collisions
Thank You Contact: Stevanus.Tjandra@edmonton.ca