Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
28 C h a p t e r 5 The RID includes two broad types of data: (1) new data col- lected and quality assured by SHRP 2 under the mobile data collection project discussed in previous sections of this report and (2) existing data acquired from public and private sources. The existing data include roadway inventory data from state DOT files, as well as HPMS, Federal Railroad Administration (FRA) grade crossings, and supplemental data, such as traffic, weather, and crashes. This chapter discusses the acquisition of existing and supplemental data. Figure 5.1 illustrates the varied data sources. The original research approach proposed development of a standardized data assessment matrix for this effort. This was completed, in part, for the larger data providers (or resources), typically state DOTs. Local agency data were evaluated a bit more broadly. An overarching consideration in the data eval- uation was if, or how, existing state and local roadway data may be used. The existing data evaluation was done in paral- lel with the RID design and helped inform the RID design on the following issues: ⢠Is a consistent level of accuracy, coverage, and feature resolu- tion for the roadway centerline framework required across all six sites? Alternatively, is the best available centerline data from each site sufficient? Consistency across all six sites would most likely require obtaining commercial road- way data. ⢠What is the minimum level of positional accuracy that can be accepted to meet the needs of researchers who will be using the roadway data? ⢠Will researchers be expected to use linear referencing methods to link attribute data to the roadway geometry? Alternatively, should the roadway database incorporate this linkage implicitly? ⢠If linear referencing methods are employed to link attri- butes to the roadway geometry, are multiple LRS, developed by the state or local agencies, used? Alternatively, is a single, consistent LRS developed for use at all six sites? ⢠How should roadway geometry features, such as horizon- tal curves or grades, be defined? ⢠Should the centerline geometry collected by the mobile data collection vendor be integrated into the roadway databases for each site? 5.1 existing Data In conjunction with the mobile data collection and supple- mental data effort, the project team also determined the char- acteristics, availability, and suitability of existing roadway inventory data and useful spatial representations for each of the six study site areas. 5.1.1 Identify Possible Data Providers SHRP 2 and the research team contacted data providers at each site, usually a state DOT, and introduced the NDS and S04 projects. The research team developed a simplified data discovery worksheet (Figure 5.2), based on the SHRP 2 Prioritized Rodeo Data Elements Lists (Assets, Geometric Features, Intersec- tions, Pavement Conditions, Roadway Inventory), and shared this worksheet with the contacts made at each site. The work- sheet served as a means to initially assess the availability and extent of various roadway characteristics, as well as those pos- sessing these data. It was separated into three primary areas: (1) roadway-attribute data of interest, (2) roadway system and extent, and (3) contact information, used to identify data pro- viders within the DOT and other agencies. 5.1.2 Conduct Site Visits After making initial contact with data providers, SHRP 2 and the research team conducted site visits to further discuss the NDS and S04 projects. Participants included state DOT staff and/or staff from other agencies, such as cities, counties, and Existing Data Acquisition
29 metropolitan planning organizations (MPOs). During these visits, an attempt was made to obtain additional information about available resources and data providers. Ideally, a pri- mary contact at each agency was established during these meetings to coordinate future requests and activities. SHRP 2 and the research team also attempted to obtain additional contacts, or data providers, through several means: (1) from those attending a site visit meeting, (2) by interact- ing with those who were not in attendance but were provided as a contact, and (3) through Internet searches. An attempt was made to obtain additional information regarding data sources and availability from these new contacts. 5.1.3 Evaluate Data and Metadata Following the site visits, the research team tried to obtain and evaluate as much metadata (captured from data dictionaries) and attribute data as possible from pertinent agencies at each site. Sometimes this required contacting agencies that had not been previously involved in the data discovery process. Given these considerations, the data evaluation process was loosely based on the crash data improvement program (CDIP) guide (http://safety.fhwa.dot.gov/cdip/). This guide was established to assist in identifying, defining, and mea- suring the characteristics of data quality within state crash databases. However, the research team felt that it was exten- sible to this effort, focusing on the six data quality charac- teristics of timeliness, consistency, completeness, accuracy, accessibility, and integration. A primary challenge of the evaluation was that data must be evaluated not only indepen- dently but also within each site and among all sites. This being considered, the research team first had to determine whether possible roadway attributes of interest were available or existed within the sites and, if so, for which systems. For example, a given attribute may only be maintained on the state system or by only one county within an NDS site covering multiple counties. Additionally, since the roadway characteristics of interest had not been fully identified at the earlier stages of the project, attributes of interest were broadly defined and assumed. The following are some of the considerations and Figure 5.1. RID data sources. Figure 5.2. Data discovery worksheet.
30 challenges associated with assessing the six data quality characteristics: ⢠Timeliness. Data must be available for the NDS time frame. Data sets under development and available during the NDS time frame was taken into consideration. ⢠Consistency. Uniformity may be greatly impacted by the various standards and protocols employed by the multiple agencies involved. ⢠Completeness. As with consistency, completeness may vary not only among agencies but also among data elements within agencies. Some agencies collect and maintain data for only a portion of the public roadways (e.g., state-maintained roads, federal aid eligible). ⢠Accuracy. Are SHRP 2 mobile data collection project accu- racy requirements satisfied? Sometimes, the accuracy of the data may not be clear or well documented or may be influ- enced by several sources. ⢠Accessibility. Accessibility to researchers may be affected by data licensing and sharing agreements. ⢠Integration. Integration describes the feasibility and/or ease of linking or integrating potentially disparate data sets with the mobile data, the NDS data, and other pertinent data available within and among the sites. 5.1.4 Findings Site or DOT visits were conducted in five of the six states: New York, Pennsylvania, North Carolina, Indiana, and Washington. A webinar was held with state DOT staff in Florida. State DOT staffs were primarily involved in meetings in New York, North Carolina, and Washington. Both state and local representatives participated in the Pennsylvania and Indiana meetings. A second meeting in New York, predomi- nantly involving local staff, was held at the Greater Buffalo- Niagara Regional Transportation Council. A second meeting was also held at the Centre County MPO, Pennsylvania, involv- ing MPO staff. Through these meetings, and additional networking and discovery, more than 130 individuals participated, were iden- tified as contacts, and/or were contacted during the data dis- covery task. Table 5.1 presents the number of contacts for each state. Data discovery worksheets were shared with meeting par- ticipants and sent to individuals or agencies identified as pos- sible data providers, primarily local agencies. A description of the NDS accompanied the data discovery worksheets. Unfor- tunately, the response rate was not particularly high with local agencies. Some of the originally proposed evaluation strategies were determined to be more applicable to database design and specifications and potentially excessive, given preliminary findings in working with some agencies. States would often request a list of specific attributes needed for NDS, as well as details about these attributes. This could not be provided at the time, because the attribute list was preliminary, and the data discovery worksheet was purposely vague in this regard. Additional challenges were presented when coordinating with multiple offices or data providers within the same agency. Because the specific data needs were not known, it was imprac- tical to investigate all the nuances of each of the data sets of possible interest. Since agencies, particularly state DOTs, provided various levels of data or metadata, additional emphasis was placed on evaluation of publicly available data. Given the existing avail- ability of these data, limited challenges should exist in mak- ing such data available to researchers, and, therefore, they may be viewed as the minimum available. In other words, existing data sharing standards and procedures are already established. In some cases, these data sets represent much of the readily usable, internally maintained data. An assess- ment matrix was prepared for the aforementioned state data. It focused on the following elements: ⢠Area. Total area (in square miles) for all counties included in the study area. ⢠Total centerline miles. Total centerline miles of all (public) roads within the study area (obtained from either Topo- logically Integrated Geographic Encoding and Referencing [TIGER] or a statewide all-roads data source). ⢠Average road density. Total centerline miles/area. ⢠Roadway centerline data source. Name and URL for the road- way centerline databases. ⢠Total centerline miles in study area. Total centerline miles included in the roadway database for all counties in the study area. Table 5.1. Number of Individuals Contacted for Data Discovery by State State Number of Contacts New York 23 Pennsylvania 37 North Carolina 16 Indiana 11 Florida 20 Washington 24
31 ⢠Percentage of centerline miles represented. Total centerline miles included in the roadway database/total centerline miles of all roads in the study area. ⢠Included road classifications. Road functional classes included in (or excluded from) the database. ⢠Source of line geometry. Primary source of digitized line geometry (e.g., 1:24,000 maps, orthoimagery, kinematic GPS). ⢠Positional accuracy. Estimate of horizontal positional accu- racy, based on data sources. ⢠Update frequency. Frequency of release of published versions of the database (from metadata). ⢠Last update. Date of most recent update (from metadata). ⢠Routable network. Initial qualitative assessment regarding whether the database incorporates network topology. ⢠Database format. Available GIS data format (typically Esri shapefile). 5.1.5 Existing Data Elements in the RID The final data elements included were selected mostly by pro- viding any available data that satisfied data quality requirements to transportation safety researchers. Overall, the variability and coverage of data items varied from state to state, and some were available for only one or few states. However, the research team did not want to limit the data set by trying to provide a common list among states; therefore, a data dump from the state DOTs was obtained. In addition to data resources from state transpor- tation agencies, Esri StreetMap Premium for ArcGIS was also included in the RID (e.g., roadway attributes include street name, functional class, speed limit, number of lanes, direction of travel). Table 5.2 presents a cross reference of mobile data items collected to available state information. The information indicates that a majority of S04B data items were not available in the existing data resources. When data items were avail- able, such as shoulder type or median presence, the data were available only for state-maintained roads. Table 5.3 lists general data items besides mobile van data that were identified in state data resources and included in the RID. Again, coverage and availability of a particular data item among states vary, but any data item that could potentially be useful for future safety research was therefore kept in the database. A broad listing of data elements from state transportation agencies included in the RID is as follows (the coverage is typically state-maintained roads): ⢠Florida: 2010 Florida Traffic Information and Highway Data (extensive database including access control type, access management, bridges, facility crossing names, functional classification, interchanges, intersections, maximum speed limits, median type, median width, number of lanes, pave- ment conditions, railroad crossings, rest areas/welcome areas, roads with local names, road status, surface width, annual average daily traffic, portable traffic monitoring sites, telemetered traffic monitoring sites, traffic signal locations, truck traffic volume, weigh-in-motion locations), and Pasco County sign inventory. ⢠Indiana: Bridges, traffic count station averages, friction, pavement condition, curves, functional classification, inter- national roughness index, lane information, shoulder infor- mation, median information, speed limits, texture, traffic section, truck traffic volume. ⢠North Carolina: Roadway characteristics, pavement condi- tion, electrical service points, signal locations, traffic vol- ume data. ⢠New York: Roadway inventory system, structures (bridges, culverts), pavement information. ⢠Pennsylvania: Bridges, intersections, pavement informa- tion, traffic information, sign inventory, roadway charac- teristics, intersections. ⢠Washington: Roadway characteristics, pavement markings, alignment, rumble strips, roadside inventory, freight goods transportation system, traffic information, bridge, culverts. 5.2 Supplemental Data For the purpose of this report, the term âsupplementalâ refers to any data item that characterizes a roadway segment that is not included as part of the mobile data collection undertaken by SHRP 2 or existing roadway data being acquired from transpor- tation agencies within the six NDS sites. Many of these supple- mental data items can be considered basic descriptive elements in transportation analysis and as such are of considerable inter- est to the transportation research community and the SHRP 2 Safety and Reliability programs. The RID includes a number of features that physically describe the roadway setting (e.g., curvature, grade, cross slope, lanes). There are, however, a number of other roadway variables that are critical to further characterizing a roadway or analyzing the operation of that roadway segment. As part of the supplemental data effort, both the SHRP 2 Safety and Reliability programs had overlapping data needs that were considered in determining the data to acquire, though the safety program requirements were the primary priority. For example, the exposure associated with a roadway segment is a common component of safety and reliability analysis. Exposure can include traffic composition and volume levels. Drivers are also impacted by changes in operations due to the impacts of seasonal weather, major public events, operating and posted travel speeds, and enforcement. System reliability is regularly
32 Table 5.2. Mapping of Mobile Data Items versus Available State Information Florida Washington Pennsylvania North Carolina New York Indiana T ie r 1 A lig nm en t Tangent State+, PI, central angle, compass bearing, degree of curve State No No No On-system only by curve class and location C ur ve Radius State No No No Length State No No No PC Lat State No No No Long State No No No PT Lat State No No No Long State No No No Direction State+ State No No No Grade Grade classes for HPMS only State No No No No Cross slope Horizontal HPMS segments only No No No No No Superelevation No State No No No No Lane Number State+ for through lanes State State+ State+ State+ State Width Total through lanes surface width State Derived State+ Pavement width, state+ Yes Type Through (state+) auxiliary (only state) State State+ State+ No Auxiliary lane incomplete Shoulder Type State highway sys- tem, HPMS, SIS State State+ State+ State+ State Width if paved State highway sys- tem, HPMS, SIS State State+ State+ State+ State Signs Lat Pasco County No State+ No No No Long Pasco County No State+ No No No Message No No Yes No No No MUTCD Code Pasco County No No No No No Number of signs on post Pasco County No No No No No Intersection Location State and county roads State State+ No No No T ie r 2 Intersection Number of approaches No State From segments No No No Control type No State No No No Only signalized state Lighting No No No No No No Median Presence State+ State Yes State+ State+ State+ Type State+ State Yes State+ State+ State+, different classification of median types Rumble strip Presence Limited outside shoulder State No No No No Location Limited outside shoulder State No No No No Note: State+ = data available on all state roads and some additional roads.
33 Table 5.3. Existing Data Items from the States, Excluding Mobile Data Items Florida Washington Pennsylvania North Carolina New York Indiana Signals State State State State, partial No State, partial Average annual daily traffic (AADT) State (also truck volume) State (also truck volume) State (also truck volume) All highways functionally classified above local State, AADT, DDHV, volume/ capacity ratio, % trucks State, section averages, also truck Pavement condition Very poor, poor, fair, good, very good (all principal arte- rial roads and some additional) Very poor, poor, fair, good, very good (state) No Alligator cracking, transverse cracking, rut, ride, patch, faulting (state) Pavement score, IRI, rutting, dominant dis- tress, overlay, crack sealing (state) Friction, IRI, and texture (on interstate roads) Friction No State State No No State Bridges, culverts Bridges, begin/end, FHWA# Bridges, begin/ end, FHWA# Bridges, begin/ end, FHWA# No Bridges, culverts, shape file (bridge charac- teristics, condition) No Posted speed limit State State State State State State Pavement surface type State State State State State Major collector and above (state), partial Functional classification State+ State+ State State State State Access/access control All roads function- ally classified higher than local State State State State No Interchanges Principal arterials, location and type Under crossings Yes No No No Intersections State, location, surface and description State, control type, illumination, and configuration State Number in a segment No Signalized, state Railroad crossings Yes Yes Yes Yes Yes Yes Rest areas Yes Yes No No No No Vertical alignment Only adequacy, HPMS segments State No No No No Terrain HPMS segments Yes No Yes Yes No Tunnels No Yes No No No No Note: IRI = international roughness index; state+ = data available on all state roads and some additional roads; HPMS = Highway Performance Monitoring System.
34 impacted by the frequency of roadway disruptions due to construction, closures, incidents, or other capacity reduction factors. The supplemental data will aid researchers in con- ducting a more comprehensive analysis of driver, vehicle, and roadway interactions and full utilization of the NDS and RID databases. These data will also allow for conducting before- after evaluations of new safety programs, initiatives, and laws introduced or changed during the period of the SHRP 2 pro- gram in each NDS site. The supplemental data effort initiated with a feasibility study before SHRP 2 gave permission to proceed with data acquisi- tion. The feasibility study addressed the availability and cost of acquiring data items specific to the SHRP 2 Safety and Reli- ability focus areas and integrating these data into the SHRP 2 RID (see Appendix D for details). An initial list of more than 30 potential supplemental data items was expanded to include roughly 125 items within four categories, including safety, reliability, common items to both safety and reliability, and policy. Input from both safety and reliability professionals helped to narrow this comprehensive list down to approxi- mately 32 items. Using this prioritized list as a starting point, the research team contacted agency staff within the six NDS sites and private data sources and recommended 20 data items for acquisition across the six NDS sites. Table 5.4 presents the top-priority data items acquired by the project team that are included in the RID. Figure 5.3 pro- vides a description of these data. In this figure, each data item is described with the spatial coverage, frequency of data sets, and time of coverage. Table 5.4. Data Items for Supplemental Data Priority Item Category 1 Crash data Common Traffic information: AADT Safety Aerial imagery Common Speed limit data Safety Speed limit laws Policy Cell phone and text messaging laws Policy Automated enforcement laws Policy Alcohol-impaired and drugged drivers laws Policy Graduated driver licensing (GDL) laws Policy State motorcycle helmet use laws Policy Seat-belt use laws Policy Local climatological data (LCD) from NOAA Common Cooperative weather observer/other sources Common 2 Winter road conditions (from DOTs) Common Work zone Common 511 information Common Traffic data: continuous counts (annual traffic report [ATR]) Safety Traffic data: short duration counts Safety Changes to existing infrastructure condition Safety Roadway capacity improvements Common
35 # Item Priority Category Description Data Type Data Frequency Time Period of Interest 1 Crash Data Common Stateâs official crash file. GIS State/County Yearly 2006 2013 2 Traffic Information - AADT Safety Computed AADT by site or segment. GIS Statewide Yearly 2011 2013 3 Aerial Imagery Common Aerial imagery at 1 meter resolution which can be associated with Geographic Information Systems (GIS). 1 Meter Resoultion Image County Based Single Year Between 2011 2013 4 Speed Limit Data Safety Posted speed limit by location, and if available advisory and special restrictions along with changes during the SHRP2 NDS period GIS Statewide Single Year Between 2011 2013 5 Speed Limit Laws Policy Any increase or decrease to the maximum statutory speed limit over the two-year duration on state roads. State Summary Document Yearly 2011 2013 6 Cell phone and text messaging laws Policy Laws which prohibit hand-held cell phone use and/or texting while driving. State Summary Document Yearly 2011 2013 7 Automated enforcement laws Policy Laws which enable or prohibit the use of automated speed enforcement, red light running cameras, or other devices. State Summary Document Yearly 2011 2013 8 Alcohol-Impaired and Drugged Drivers laws Policy Laws enacted that prohibit driving while impaired, publicizing and enforcing those laws, and punishing the offenders. State Summary Document Yearly 2011 2013 9 Graduated driver licensing (GDL) laws Policy Laws specific to GDL (nighttime hours, number of passengers, cell phones, learners permit, seat-belt use, etc). State Summary Document Yearly 2011 2013 10 State motor cycle helmet use laws Policy Laws covering helmet use for both driver and rider and if there are any differences by specified ages. State Summary Document Yearly 2011 2013 11 Seat Belt Use laws Policy Primary enforcement belt use laws permit seat belt use law violators to be stopped and cited independently of any other traffic behavior. Secondary enforcement laws allow violators to be cited only after they first have been stopped for some other traffic violation. State Summary Document Yearly 2011 2013 12 Local Climatological Data (LCD) NOAA Common A monthly NOAA summary from major airport weather stations that includes a daily account of temperature extremes, degree days, precipitation and winds. Also included are the hourly precipitation and abbreviated 3-hourly weather observations. Access Database by Station Monthly 2011 2013 13 Cooperative Weather Observer/Other Sources Common Hourly precipitation amounts obtained from recording rain gages located at National Weather Service, FAA, and cooperative observer stations which includes inches to tenths at local standard time; includes maximum precipitation for nine time periods for some stations; these data are final quality controlled copy and have a 4 to 6 month time lag. National Climactic Data Center Storm Event data which includes storm events by county including tornadoes, thunderstorm winds, and hail. Clarus data where available. Weather Underground (personal weather station information) data if available including (daily high/low/avg; Temp(F), Dew Point(F), Humidity(%), Sea Level Pressure (in), Visibility (mi), Wind (mph), Precip (in). Access Database by Station Monthly 2011 2013 14 Winter Road Conditions (DOT) Common Winter Travel Advisory by road segment and time. When available this will include roadway and segment limits; Roadway Status (clear, wet, snow, ice, severe, closed); Pavement Condition; Weather Condition; and date/time of last update GIS Statewide Yearly 2011 2013 15 Work Zone Common Data including location, limits, time period, and if available the impact to travel (minimal, moderate, severe). GIS Statewide Yearly 2011 2013 16 511 Information Common Includes Incidents/closures; Weather/Alerts; Winter Weather Operations and Advisories; Special Events; Bridge and Tunnel Status; Construction; Travel Speeds (e.g. <30mph, 30-49mph, >50mph) where available. GIS Statewide Yearly 2011 2013 17 Traffic Data-Continuous Counts (ATR) Safety Continuous count data from permanent equipment which operates 365 days a year. GIS Statewide Yearly 2011 2013 18 Traffic Data-Short Duration Counts Safety Short count data from portable equipment which typically operates from 2 days to 7 days. These counts usually vary in frequency per location from every year to up to 5 years in between counts. GIS Statewide Yearly 2011 2013 19 Changes to existing infrastructure condition Safety Data documenting changes to infrastructure condition such as overlays or replacements that improve condition without adding capacity or changing geometry. GIS Statewide Yearly 2011 2013 20 Roadway Capacity Improvements Common Data documenting major operational and capacity enhancements such as additional lanes, intersection treatments, traffic control, traffic signalization, reversible lanes, designated lanes, roundabouts, etc. GIS Statewide Yearly 2011 2013 1 2 Figure 5.3. Supplemental data collection description.
