Hello forum,
I'm trying to find the best position from about two months recording of GPS coordinates with mobile phones (so the file is huge)
What I am trying to do is to pick out the records (column D) with the best group of cells within 100 meters based on ItemID (column C) just to get rid of large errors in GPS positions. I have sorted column C and D in ascending order and from this try to find the group of values that fit best for the 100 meter value.
I tried this formulabut this works only if the first value always is correct and that is not always the case as the GPS signal can drift a lot (bad weather, signal-shadow, battery status, etc). The distance in meters is calculated from where the point is supposed to be and with this data I'm trying to see if I need to update with better coordinates later, using some sort of mean or median number. There can be big drifts within each itemID, it can be only a few that is within 100 meters, perhaps when there is a very few (< 10) in that case if possible extend the value to 200 meter at is most or else all from this ItemID should have errors and be ignored in my next step. I hope I explained well enough, the attached file is sample from the data I am using and the validation column (column E) is from the code above I been using that does not work for my purpose.Please Login or Register to view this content.
Thanks
Bookmarks