Archived Data User Service (ADUS):
Workshop #131
Sunday, January 12, 2003
Transportation Research Board 82nd Annual Meeting
James Pol, Federal Highway Administration
Ohio & Kentucky
Approach to Data Archiving in Cincinnati
David Gardner, Ohio DOT & Rob Bostrom,
Kentucky Transportation Cabinet
Developing Efficient Data
Archive Designs for State of Virginia
Simona Babiceanu, University of Virginia
Data Archive Designs Based
on Traffic Monitoring System Experience
Joe Wilkinson, Chaparral Systems
Data Archive in Minnesota
using Common Data Format (CDF)
Taek Mu Kwon, University of Minnesota-Duluth
11:20 – 12:00 ROUNDTABLE
DISCUSSION
The following general issues (with associated participant comments) were discussed during the roundtable portion of the workshop.
Challenge of measuring classification using loops in an urban environment (i.e. short headways)
o Kentucky experience has been mixed. Have relied on manual classification count to check/calibrate the data
o Washington
– difficulty is primarily the inconsistency. Some
days are good, some are very bad.
University involvement in ITS Data Archives
o Pravin Varaiya, UC Berkeley offered an overview of the PeMS system. PeMS started in 1998 as a research project. Morphed into a semi-production project. Production version will run in CalTrans & a duplicate version will continue to run in the university. $2M cost of system ($0.5M per year). 2-3 terrabytes of data available on-line. 2 gig per day added. Very small % of users wish to directly query db. PeMS website serves an average of 4,000 hits per day. Big uses – graphs & contour plots. All data collected by PeMS is tested with QC tests, if fail, data is imputed. But – raw data is still always retained. Development has involved 2 professionals + 3 graduate students. Production software developed by a private company.
o Brian Smith, University of Virginia provided an overview of the Virginia Smart Travel Lab experience. Stated that close partnership of VDOT + university was critical to making the data archive work in university setting. Discussed concerns regarding stability of funding (i.e. not only research
o
The group discussed challenges associated with funding applied vs.
theoretical research.
Data Archiving Costs
o Administration and maintenance costs are critical components
o Storage
costs are relatively minor
Participants agreed that it is insufficient to only talk about archiving data – real issue is how to retrieve and use the data.
There was consensus that a need in the community is to clearly document who archived data users are and how they are using the data.
o
Most operators have never had access to this data. So
– they really have no idea as how to use it. When
you’re trying to assess user requirements – need to be creative in
extracting needs from users. They won’t just
have a list to hand you. User community may
be classified as “operations planning”
Draft of
ASTM Archived Data User Service guide for archiving and retrieving
ITS data will be distributed for final ballot (ASTM membership +
comments from general public). Reviews &
reactions would be very much appreciated. This
is the first data archiving standard developed as part of the
standard development process. Will next be
tackling individual types of data (flow, incident, …). Current
standard addresses metadata fully.
Consideration for Future Workshops
o Data quality
o Who are the users and how are they using it?
o How do archives support measuring the performance of the transportation system?
o Comparison of different archival techniques
o How to build systems efficiently?
o What can you do with this data – sample applications.
o
Focus more on workshop/discussion – not as much emphasis on
presentation