NATMEC2016 - Development of QAQC Processes for BikePed Data
- 1. Reporting / Analysis •Data aggregated to station level •Released annually (12-month reporting period) •NMVDP Format reported to project stakeholders •TMG Format exported to FHWA TMAS system •Data used for analysis: • Hourly, Daily, Monthly, and Seasonal Patterns • Factor and Volume Groupings •Statistics summarized using AASHTO Method: • Annual Average Daily Bicycle Traffic (AADBT) • Annual Average Daily Pedestrian Traffic (AADPT) • Assess site characteristics • Predict activity patterns • Minimize site-specific data collection issues Pre-Installation Site Selection Virtual and In-field Site Inspection Short Duration Count Procurement Station Layout Diagrams Equipment Configurations Matched to Station EquipmentSet-Up Installation Data Gathering for Onboarding Equipment Installation Procedures Equipment Testing and Troubleshooting Onboarding Equipment Inventory Installed Equipment Diagrams Station ID’s Defined Data Streams for Download Phase I of North Carolina’s Non-Motorized Volume Data Program (NMVDP) was conducted in NCDOT Divisions 7 and 9 in the Triad/Piedmont region of NC. Continuous Count Stations (CCS) to monitor bicyclist and pedestrian traffic at twelve locations went live in late 2014. These stations cover a mix of sites across different land uses, travel patterns, and volume groups. The following programmatic elements (indicated in RED in the diagram) were piloted to select, install, and ensure data quality for the twelve CCS stations. Introduction Kristy Jackson Research Associate knjackso@ncsu.edu Sarah Searcy Research Assistant sesearcy@ncsu.edu The authors would like to thank the project manager Sarah O’Brien from ITRE and Lauren Blackburn from the Division of Bicycle and Pedestrian Transportation at NCDOT for their oversight and guidance of this project. Acknowledgements • Ensure quality data streams • Provide accurate labeling and metadata • Minimize installation errors • Ensure functional equipment at onset • Document inputs for onboarding station Training for Program Partners Ownership and Maintenance Agreements Coordination of Roles and Responsibilities On-Site Routine Maintenance Training •INVALID data examined to determine potential equipment malfunctions •Maintenance is scheduled in timely manner •Maintenance may occur as a result of findings of equipment validation process •Revalidation occurs if necessary itre.ncsu.edu/focus/bike-ped D E V E L O P M E N T O F Q A / Q C P R O C E S S E S FO R B I C YC L E A N D P E D EST R I A N DATA AgencyCoordination EquipmentValidationDataHandling • Data is downloaded quarterly (Phase I protocol) • As collected from Continuous Count Station (CCS) • Minimal data manipulation occurs to prepare data for QA/QC checks QA/QC Checks Applied : Gap Check Consecutive Zero Directional Distribution Range Check Checked • Data is checked quarterly • Data is flagged for potential equipment malfunctions • Processes are automated to the extent possible • Flagged data is reviewed and marked as VALID, ATYPICAL, or INVALID • INVALID data may initiate maintenance Cleaned • Data is cleaned quarterly • INVALID data days are removed from dataset • Cleaning protocols followed • Processes are automated to the extent possible Corrected • Aggregated to daily data with hourly intervals by mode and site • Application of correction factors from validation process • Determine corrected Directional Distribution Raw Data Raw Data • Ground-truthing of raw data • Video data collection • Accuracy analysis through comparison of automated counts to manual counts EquipmentMaintenance • Balance budgetary constraints and program outcomes The accuracy of non-motorized data collected by Continuous Count Stations (CCS) is important to be able to create sound estimates of walking and bicycling volumes and factor data from Short Duration Count (SDC) locations. Data quality is important for any application because it affects the credibility and usability of the data for agency decisions. In a volume data program, there are numerous points at which Quality Assurance and Quality Control (QA/QC) processes or procedures can be applied before, during, and after data is collected. The diagram shows the actions taken to ensure data quality for the NMVDP. Methods Selected sites were installed with assistance from local agency staff in the following NC municipalities: Carrboro, Chapel Hill, Greensboro, and Winston-Salem. Data was monitored for the 12-month reporting period and invalid days were removed from the data. The equipment at each CCS underwent a validation process to ground-truth each stream of non-motorized count data and correct it for errors related to data collection. A data summary is provided for each station based on the resulting count data, with days of missing data and data related to equipment errors removed. Results NMVDP Naming Format: Mode City Side of Street Direction of Travel Location Bicycle and Pedestrian Program Institute for Transportation Research & Education North Carolina State University N O R T H C A R O L I N A N O N - M O T O R I Z E D V O L U M E D A T A P R O G R A M Table 1 Summary Volume Statistics (1) (2) (3) Durham, NC – American Tobacco Trail (12/01/2014 – 11/30/2015) Pedestrians Highest Volume Lowest Volume Season Spring Winter Month May February Day of Week Saturday Friday Date Sun. Feb 8, 2015 (1,056) NA Peak Period Weekends 8AM-5PM 12 Month Pedestrian Count 101,720 Annual Average Daily Pedestrian Traffic 349 AADPT Bicyclists Highest Volume Lowest Volume Season Summer Winter Month August February Day of Week Sunday Thursday Date Mon. May 25, 2015 (1,038) NA Peak Period Sundays 10AM-4PM 120 128 12 Month Bicycle Count 92,480 128Annual Average Daily Bicycle Traffic 260 AADBT North American Travel Monitoring Exposition and Conference Bike and Pedestrian Data: Data Wrangling and Access May 1, 2016 - Miami, Florida