Spurious data in Ashton Park Go Active results

[Siequnu]

The headline stats appear to be wrong. The Go Active mode results in a large decrease in active commuters, 0% walking and 12% more driving! Data being pulled from all-sites.geojson. The chart data, which is generated from the mode-split.csv doesn't match, but it looks to be more realistic.

I had a look at other sites but couldn't find any apparent mismatches, but it does raise the question that clearly the mode-split and all-sites are now out of sync for at least one site — are these being generated together or separately?