Data Quality Checks and Validation

The quality of WIM data is an important issue.

It is recommended to develop data verification procedures that can be implemented by the party responsible for the data collection and uploading into a corporate database.

The data owner should also perform random verification of the data before its release.

Some of the basic data anomalies to check for include:

  • missing data (the system has not recorded any data over a certain time interval)
  • data is very different to historical data at that site (the cause need to be identified, sometimes data differences are genuine, e.g. due to significant changes in freight patterns)
  • excessive overloading.

A manufacturer usually provides diagnostic software for specific WIM equipment. It also provides vendor validity codes to determine sensor performance and sensor degradation.

Validity codes are assigned to vehicle record events where there was a problem with sensor detection/failure, acceleration, straddling, minimum speed, classification, or sensor count mismatch, etc.

Further, data owners should develop metadata statements to accompany the release of any WIM data. Metadata is information about data and allows data users to assess the relevance of the data for their need.

Download an example of metadata provided by Main Roads Western Australia .

 


More in this category: WIM Data Output »
back to top