Hi there, I have 234 polylines (which represent 234 baboon travel days) and I need to create a "habitual route network" to determine whether the baboons travelled through a "network" of habitual used "routes". To do so, I would like to identify common (segments of) paths (at least 100m in length and no more than 50m apart) that were repeated 1) at least four times and 2) at least 2 times. All the averaging of lines (eg collapse dual lines), is based on only 2 lines and I'm unsure how to do this for 234 lines. I don't know whether there are any solutions in the network analyst, but hoping somebody may be able to point me in the right direction. Many thanks for reading this and for your advice. Louise
Interesting 🙂 The criteria is too rigid to ask baboons to follow 🙂
i am assuming, you generated the lines by connecting GPS waypoints of the same date. there are a couple of ways to do this. The difficult criteria to achieve is a minimum of 100m length .You can get exactly what you are looking for if you have programming support.
To start with the simplest solution, first, break the line feature at each waypoint, ie., line is now segments between two waypoints on a line. Select the layer and add a field named "date", as text field and give it a value 23 which means 23rd of a month (for all lines ofcourse). Now "save as" this file with a different name. Add the two layers to the map window. Now Select one of the layers to do a spatial join with the closest option with a search radius of 50m. Use one to one and define the merge rule as "join" on field "date" with comma delimiter. This operation will help you to avoid duplicate counts. After your finish this spatial join operation, the resulting layer will have a "join count" indicating the number of line features found within 50m, but this count can have lot of duplication, so use the merged "date" values to correct this error (may be in excel).
this whole operation will give you the paths that were travelled 2 times or 4 times, but not if they overlap 100m. Though not entirely correct, but you can roughly say the overlap distance is same as the length of the segment under observation. To do the overlap 100m criteria, you will need to create buffers and iterate almost the same procedure as above for each buffer region with a programming code. But it is possible to answer your original question.
For baboons as in question, I might use different criteria, for example, paths proximity to food source, shade, water, divide the whole area into grid and calculate the time spent in each grid cell and possibly identify sources of long durations etc, Nevertheless, if you have programming support, your task will be easier,
I am not sure if a solution is possible with network analyst. May be the thoughts above can bring a solution from others.