Traffic flow indicator has been reintroduced

More than half of sites had no data

We removed access to our Traffic Flows indicator in the December 2022 refresh of the QEM, published in February 2023. This change followed our review of input data as part of the QEM refresh round, when we realised that, all of a sudden from the October 2022 month onwards, the number of traffic sites with reported monthly data had dropped to below half the previous number. 

As Chart 1 outlines, prior to October 2022, more than 80% of sites had reported data (since we started tracking this measure in 2013/14 – except for one month in 2018). Then in October 2022 the proportion of sites reporting fell to just 28%, and then didn’t make it back to above 60% at a sustained rate until mid-2023. 

Enquiries with NZTA highlighted a few reasons for the poor level of data availability. These reasons included aged data collection hardware at our sites, as well as resurfacing activities over summer damaging inductive loops/traffic counting sensors. 

However, the most important change, that coincided with the drop off in data availability, appears to have been a change of NZTA’s national telemetry contractor. There were apparently some considerable knowledge gaps in the handover process, which led to the non-reporting or lower data availability of site-level traffic flows. 

Data availability improved to above 70% by September 2023, and has remained above that level since then. In August and September 2024, traffic flows reporting for National Telemetry Sites moved above 80% for the first time since October 2022. 

Filling in the data gaps

The consequence of missing traffic data is considerable – our Traffic Flows indicator in the QEM requires local-level data, and if a whole lot of sites aren’t reporting, we can’t construct the traffic flows indicator. 

Although reporting outcomes have improved, we were still left with over 1,000 individual missing observations. Over the last few months, we have used a range of approaches to plug the data gaps. 

To interpolate these missing values – essentially inserting a figure for a missing value – we looked to a few options. 

• For some missing values, we examined the trend in traffic flows in neighbouring sites that had reported data, and used that trend to approximate what the missing data likely looked like.
• For other missing values, we examined how traffic flows data trend in relation to other sites in the local area, to determine how traffic flows in that area likely performed over time and used that analysis to approximate what the missing data likely looked like.
• Occasionally, where there were a number of missing data points for a number of related traffic sites, and no other approaches were viable, we used previous traffic flows for that area to determine the usual monthly pattern for traffic flows, and applied that pattern to the traffic flows for a site. 

A health warning for Oct 2022 to Sep 2024

These various approaches mean we have now plugged all 1,049 gaps in the data, and we have a viable set of data to base our Traffic Flows indicator off. With more than 70% of sites reporting since September 2023, the Traffic Flows indicator can be treated as a reliable indicator for recent quarters, including September 2024. 

However, given the high degree of interpolation of missing data we had to do over the two years between October 2022 and September 2024, data over this period should be treated with caution. 

For any questions about the Traffic Flows indicator, please contact us.