Announcement

Collapse
No announcement yet.

Questionable accuracy on the feed-stats page

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Questionable accuracy on the feed-stats page

    I decided to ensure that I was logging the aircraft I saw into a database (virtual radar is super handy with the plugin for this). I then ran my own numbers from the same data.

    I can't quite replicate your results (There's a question as to if you're using first contact range or last contact range (or maximum contact range?) on the distances), but I have huge disparity between your data and what I've got for my receiver:

    Take the 3rd of November 2018:

    Criteria Your data (as per the FR24 site) My data
    Aircraft seen 1245 1972 (unique aircraft), 3561 total
    Positions reported 2353 2682
    Maximum distance 1 155nm 206.72nm
    Average distance 2 ~80nm 92.76nm
    Hits reported 3 11920 550426
    Histogram
    • <50: 601
    • 50-100: 652
    • 100-150: 147
    • 150-200: 5
    • 200>: 0
    • <50: 212
    • 50-100: 1507
    • 100-150: 635
    • 150-200: 320
    • 200>: 8
    Footnotes:

    1: Do you count just first, just last, maximum distance of the aircraft? Do you use Haversine formula for the great circle distance?
    2: The average per-day distance isn't listed, but this is the average as listed on the share-statistics page
    3: Your page defines a hit as "number of updates from aircraft to receiver - regardless of update content". In my figure, I've only included the positional update counts. (if I counted all the updates it'd be a much larger difference, in the order of 2-3 million)

    If I could post links then I'd also include screenshots; the polar plots are vastly different, and I've got a vastly different histogram and aircraft seen barchart (which I've upgraded to include the positions seen as well, which looks super nice).

    Granted, I do also see a few military aircraft (on the order of 10-15 a day), but their counts are low enough that it can't explain the discrepancy I'm seeing.

    Basically, can we have an honest chat about how you're constructing the data on the feed-stats page and why I can't seem to correlate any of your numbers for my own data? Maybe we can work together to increase the accuracy of this data and help everyone at the same time.

    (If anyone wants to see my FR24 stats, have a look at ID 18457)
Working...
X