In this high speed rail screenshot, the average speed is around 100 mph but only points under 10 mph are being plotted in the speed chart
This indicates a disconnect between the underlying data and data presentation
Good screenshot to work from — and your reading of it is accurate: the chart is only showing the final stretch of the trip. What’s going on is two different instruments rather than one broken one:
The average (125 mph) is simple arithmetic: the trip’s distance divided by its duration. It doesn’t depend on individual speed readings, so it stays right even when GPS conditions are rough.
The chart plots individual recorded speed readings — and it deliberately plots only readings that arrived with good GPS confidence. It’s not filtering on speed: the cutoff is confidence, which on this trip happened to correlate with speed — satellite reception inside a fast-moving metal carriage is heavily degraded, so the readings from the high-speed run mostly didn’t meet the bar, while the slow approach at the end (good reception) did. The chart then scales its axis to what it plotted, which is why it tops out around 7 mph.
So the underlying data is intact — the full path and distance are there, which is exactly why the average is right — and the chart is filtering on purpose, because plotting low-confidence readings would draw a confident-looking line out of unreliable numbers.
Your point that an empty stretch of chart doesn’t tell you why it’s empty is a fair one though, and it’s now filed as BIG-697 with your report as the case for it.