Hello,

If I go to My ADS-B and then look at the “Daily Median Stats (past week)” the numbers today, Saturday, are the same as the total Positions Reported/Aircraft Seen from the past Monday. Earlier this week, the numbers “Daily Median Stats (past week)” were from the past Tuesday.

Is this a bug or do I have a misunderstanding of “median” ?

Thanks, you guys do a great job, kudos to you and atta-boy.

Fuzz

It’s misunderstanding of the term median. Your weekly median will always be one of your prior days in the last week.

to calculate median:

Order all the daily positions and flights by number.

The number in the middle is your median.

K1BAA

I actually looked up the definition of median and thought that use of the word in my stats was wrong, I see now that it isn’t. Thanks

Now, once again, I am not convinced. Today, Tue, the median stats are from yesterday, no way can I postulate that yesterday is the median of the past week.

Fuzz

The last 7 days flight numbers from your stats are: 1049, 1596, 1364, 1488, 1860, 1771, 1727.
Sorting these gives: 1049, 1364, 1488, 1596, 1727, 1771, 1860
The median is the middle (4th) number: 1596.
The median shown on your stats page is 1596.

So, uh, what’s the problem?

My understanding was the problem. I did not understand that the values were sorted smallest to largest, not in order of days of the week. Got it…

Ok, no worries You know what they say about lies, damn lies, and statistics…

If you’re somewhat retentive about these statistics (as I am – and remember what analysis starts with…), you can build an Excel spreadsheet to track and mirror these numbers; Excel’s builtin median function handles the sorting and selection quite nicely. Once you’ve built that spreadsheet and start populating it with the daily numbers, you can also figure out how it’s going to change, or not change, for the next day.

I keep such a spreadsheet (surprise, surprise), tracking and comparing numbers from my production and development ADS-B boxes. I’ve also selected another PiAware site in my area, one with a history of steady statistics, for use as a reference station.

Besides calculating statistics between my own boxes, I also compute the same stats against this reference station. This helps me cancel out the effects of daily variations in the underlying numbers. Helps to answer questions like, “did that change really drop my numbers, or is it just the usual weekend variation?”

cheers–

bob k6rtm