I have gotten a fairly large polyline-dataset of GPS-recorded tracks. I'd like to have a sample along the tracks for different variables. The problem is that the data has overlapping tracks. This type of situation occures in example when same track has been walked (recorded) back and forth:


Points in the picture are 3 m apart from one another (measured along the polyline). As you can see, there are duplicate observations, which would corrupt my sample.

What is a convenient way of moving points (=observations) that are too close to each other and, most importantly, how do I keep one observation and delete the other?

I think I should clean the polyline data first and then take the sample, but I'm still figuring out how to do this.

  • 1
    Could you elaborate on the sense in which you believe you have "duplicate" observations? It is not at all evident (based on general statistical principles, anyway) that your procedure would corrupt anything. (I would guess that the person walking along this track was observing something nearby and you might be concerned about over-weighting their observations. That concern is legitimate: but it is handled through proper statistical analysis of the data, not by moving your points.) Perhaps you could tell us why you are sampling and what analyses you intend to do with these observations?
    – whuber
    Feb 7, 2014 at 16:26
  • Overweighting is exactly why I'm concerned. I'm having twice the same info.I have factorial data (track condition; good, average, damaged etc.,in the picture seen as green and yellow) covering the track length. In addition I have GIS layers such as slope, DEM, ruggedness etc., and I'd like to have sample from them, so I could take that data into R. My idea was to take the layer values from every meter or so (although here 3 m is used). I Also think this issue is interesting because it's a bit like having slices when using polygons: you are having twice the same information in your GIS data.
    – reima
    Feb 10, 2014 at 7:31
  • Just to be clear; these points are made from the polyline data. The line data itself is the observation from the field.
    – reima
    Feb 10, 2014 at 7:38


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.