美国养护工区特征分析

2.Project-specific Websites

(e.g., )

3.National Traffic and Road Closure Information PageThe Rand McNally website covers roadwork activity from major construction projects across the nation. The data from this resource has been utilized in other studies looking at the sources of traffic congestion [6]. The data provided identifies projects by location and date but does not contain details of work zone activity such as schedules of lane closures by day of week or time of day [7].

Project-specific websites contain a wealth of information regarding primarily major urban rehabilitation and construction projects. However, there were too few of these resources to provide a representative cross-section of roadwork across the nation.

In order to collect data on some details of work zone activity and to improve the breadth of our analysis, data for this effort was collected from web-based resources linked to the National Traffic and Road Closure Information page. The level of detail was somewhere between the detail of the project-specific websites and the project-level approach of the Rand McNally resource, although this differed significantly from state to state (and even within a state). Further, collectively, the state web resources covered roughly twice the number of work zone records found concurrently in the Rand McNallyresource [7].

Of the 50 state and other resources listed, data were systematically captured on individual roadway projects and their associated work zones in 13-20 states over a two-week “snapshot” period.The pilot snapshot was conducted in thirteen states, the four semi-annual snapshots were conducted in twenty states. Figure 1 shows the statesanalyzed in all five snapshot efforts. The final sample of twenty states, kept consistent across the four semi-annual snapshot efforts was selected based on the number of work zones described and the amount of detail provided for each work zone.Since the data were organized differently state-to-state, and in many cases descriptions provided in text form, an analyst to inspect and code each record was required.

Managing and interpreting data from the disparate state resources was a significant effort. In order to avoid double-counting of projects posted twice over the same time period, automated routines were developedto speed the identification and removal of duplicate records. A systematic coding rubric was required so that interpretation of various data elements was consistent regardless of the analyst performing the interpretation. Even with partial automation, data acquisition and coding was a significant and time-

consuming element of each snapshot study.

Our experience with the acquisition and coding of data elements suggests that if states followed a consistent set of guidelines for data elements and their definition,

comprehensive national studies of work zone activity would be significantly cheaper and more valuable. A recent paper explores the possibility of establishing a set of national guidelines for posting work zone information [8].