1. Linking Data with Decision Making: Data Stewardship Issues Gerardo W. Flintsch Gerardo W. Flintsch Roadway Infrastructure Group, VTTI Virginia Polytechnic Institute and State University James W. Bryant Asset Management Performance Improvement Virginia Department of Transportation

2. Overview  Introduction  Data Stewardship  TAM Survey  Findings & Recommendations  Data Collection Framework

3. Introduction  Data collection is a key part of Asset Management  Transportation agencies collect data for their assets in order to support decision making  Data collection activities do not always specifically support decision processes

4. Asset Management  Most transportation agencies around the world have begun moving towards Asset Management  Primarily by considering the integration of individual management systems  Some organizations are trying to develop comprehensive Asset Management Systems

5. Decision Levels  The levels of decision making are:  Strategic Level  Network Level  Programming  Project Selection  Project Level  Different focus/impact on the system  Relate to different decision processes

6. Decision Processes & Data Needs Strategic Level Network Level Project Level Programming Level Project Selection Level

7. DATA STEWARDSHIP

8. Data Stewardship  Process of monitoring, maintaining, and enriching the value of the organization's data assets.  A data steward is a person empowered to implement the data stewardship process Strategic Insights, Inc  The formalization of accountability for the management of organizational data. bitpipe

9. Foundational question: What to Collect?  Decision often based on  Wish list (“nice to have”)  Existing/ historical data collection processes  Can lead to data collection becoming an end in itself  Excessive or inefficient data collection could compromise project

10. What to Collect?  What Asset Management decision do we need assistance to make?  What data are needed to support those decisions?  Can we afford to collect these data initially?  Can we afford to keep the data current over a long time period?

11. TRANSPORTATION ASSET MANAGEMENT SURVEY

12. Research Objective 1. Investigate how data collection is linked with decision processes – especially at the Project Selection level 2. Documentation of “Best Practices” 3. Framework for effective and efficient data collection to support Project Selection decision

13. TAM Survey - Part 1  General Agency Information on Asset Management  Endorsement and implementation  Existing/planned individual management systems  Existing levels of decision making  Identification/ rating of decision processes  Identification/ rating of Project Selection criteria

14. TAM Survey - Part 2  Roadway Asset Management  Data management, collection methods and integration  Rationale behind existing/ planned data collection  Evaluation of roadway asset data for Project Selection  Identification of links between Project Selection and data collection activities.

15. TAM Survey - Development  Web page developed using Macromedia Dreamweaver and PhP  Web page uploaded @ www.am.vtti.vt.edu www.am.vtti.vt.edu  Database created using MySQL  Access to the survey was granted through recipients’ email addresses  Sent to 103 transportation officials in all 50 US DOTs and Puerto Rico  Open for responses for 3 weeks

16. TAM Survey - Analysis  48 responses were received from 40 states  Response rates:  78% in terms of individual states  47% in terms of individual respondents  Statistically analyzed

17. TAM Survey - Results Part 1  Implementation of Asset Management system  24 states are in still in the planning phase  11 states have already implemented one  Individual management systems  Pavements (39)  Bridges (39)  Maintenance (32)

18. TAM Survey - Results Part 1  Integration of individual systems  Most States are still in the planning phase  Most integration efforts have taken place for Pavement and Bridge management systems (20 states)

19. Defined Decision Making Levels

20. Please rate the following Asset Management decision processes in terms of their relative importance within your agency/organization.

21. Ranking of Asset Management Decision Processes Asset Management Decision ProcessesAverage Ranking Performance evaluation and monitoring3.54 Fiscal planning3.48 Project selection3.33 Resource allocations3.31 Policy formulation3.13 Program optimization and trade-offs3.10 Program delivery/ project implementation3.02 Performance-based budgeting2.96 Audit, reporting and communication2.81 Development of alternatives2.77 Impact analysis2.65 Key: 4 = Very Important, 3 = Somewhat Important, 2 = Not Very Important, 1 = Not Important at All

22. Please rate the following criteria according to their level of importance for selecting projects that are candidates for funding and implementation within your agency/organization

23. Ranking of Project Selection Criteria Project Selection CriteriaAverage Ranking Available budgets/ earmarked funds3.75 Engineering parameters3.44 Public demands/ user opinion3.27 Project significance3.27 Agency costs/ benefits 3.19 Usage of project 3.08 Environmental considerations3.06 Geographic distribution of projects/ funds2.83 User costs/ benefits 2.77 Community costs/ benefits2.65 Distribution among asset types2.46 Ease/ difficulty of implementation2.38 Proximity of project to major urban areas2.25 Key: 4 = Very Important, 3 = Somewhat Important, 2 = Not Very Important, 1 = Not Important at All

24. TAM Survey Results Part 2  Asset Management roadway database  75% have already developed one (30)  Majority of remaining states (5) are in planning phase  All agencies collect data for pavements and bridges

25. Roadway Asset Data Collection Types & Methods

26. Agency Data Collection Rationale

27. Which roadway asset data are most important in your agency’s view for the selection between two projects, e.g. between different pavement projects or between a pavement project and a bridge project?

28. Ranking of Roadway Asset Data for Project Selection Roadway Asset Data Average Ranking Structural condition3.77 Functional condition3.67 Usage3.29 Initial agency costs3.23 Life Cycle Costs2.96 Attributes/ characteristics2.90 Customer/ user feedback and complaints2.83 Location2.67 Key: 4 = Very Important, 3 = Somewhat Important, 2 = Not Very Important, 1 = Not Important at All

29. Link Between Data Collection & Project Selection 52.5% of Agencies 32.5% of Agencies

30. CASE STUDY VDOT

31. . Decision-Making Levels Identified  Strategic Level:  Decisions regarding cross-agency issues and funding allocation across the highway maintenance, mobility management, and operation programs.  Programmatic Level:  Operation of the programs (e.g., highway maintenance) to achieve established state network performance targets.  Project Level  Works related with District operations, such as identification of funding for various project

32. Core Asset Management Business Processes  Condition Assessment  Need-Based Budget (financial modeling)  Project Selection  Feedback (Web-based TOOLS)  Work Accomplished and Inventory Updates

33. Information Needs  Pavements  The pavement condition data is gathered independently from the Asset Management system  PMS group that performs condition surveys (mostly windshield survey)  Bridge  Information collected to feed PONTIS  Requested by a federal mandate (NBI)

34. Information Needs  Other Highway Assets  The information about the other highway assets (8 types of infrastructure assets) is gathered through a simple strategy that is conducted state-wide  Random Condition Assessment (RCA) collects data from the assets that are most significant to the maintenance program budget.  Pressure to increase the number of assets

35. Data Collection Lessons Learned  The agency really needs to target the data collection to the most significant assets and to the decisions that has to be made.  VDOT had a project that tried to capture every highway infrastructure asset on every highway state-wide.  The project was started in 3 counties as a pilot  It was found to be too expensive and overwhelming.  Current approach: build the inventory over time

36. Data Collection Assessment  Undergoing  How is the system put together and how it meets the need of the agency?  How well are the processes documented?  What analysis tools are available and how are they being used?  How good is the repeatability of the process?

37. Other Case Studies  Maryland  Visit conducted & documentation undergoing  Washington, DC  Scheduled  Possibly North Carolina

38. FINDINGS

39. Findings  US transportation agencies still collect data predominantly based on past practices and staff experience.  There is a trend towards following data collection standards and input needs  Most important criteria for Project Selection  Available Budgets/ Earmarked Funds  Engineering Parameters  Public Demands/ User Opinions

40. Findings  Most important data for supporting Project Selection:  Structural condition of asset  Functional condition of asset  Usage of asset

41. Findings  US Transportation agencies have explicitly defined decision levels (business processes) and are moving towards rationalization of data collection activities.  Past practices and staff experience → collection standards and input needs  Agencies have identified links between data collection and decision making at the project selection level but not all of them have explicitly documented these links

42. Recommendations  Target the data collection to the most significant assets and to the decisions that has to be made.  Collect only the data you need  Collect data to the lowest level of detail sufficient to make an appropriate decision  Collect data only when they are needed  Use pilot studies to test the appropriateness of the approach

43. DATA COLLECTION FRAMEWORK (Preliminary – Project Selection)

44. Project Selection Data Collection Framework Identification of candidate projects and corresponding treatments for all assets under consideration Identification and/or definition of data needs for the analyses/evaluations to be performed Identification of existing databases and data quality assessment Identification of missing data elements Selection of the appropriate data collection methodology and implementation FEEDBACK LOOP: Evaluation of assessment tools/ data inputs/ data collection and overall results (Feedback to upper levels) STEP 1 Identification of project evaluation/assessment tools and their requirements STEP 2 STEP 3 STEP 4 STEP 5 STEP 6

45. Many Thanks to…  Aristeidis Pantelias  All survey respondents  Federal Highway Administration  AASHTO Task Force on Protocols, Procedures, and Technology for Asset Management Condition Data Collection  VT Statistics and Survey Departments  …

46. Thank you!