Relative "curvy-ness"

Question

I work in the rail industry and part of my work involves comparing routing options. I have a number of metrics that I use but one that I would like to incorporate is the relative "curvy-ness" of one route vs another. does anybody have any insight on metrics or tools that I could look at to derive something like this?

danretired · Answer

Divide a polyline by a value—ArcGIS Pro | Documentation if you have just a few routes to analyse

or the pair of

Generate Points Along Lines (Data Management)—ArcGIS Pro | Documentation

Split Line at Point (Data Management)—ArcGIS Pro | Documentation

aburton07 · Answer

Thanks for the excellent feedback! I have looked into sinuosity in the past, but it seemed like it was more of a measure of route circuity than a measure of the 'curvy-ness' of the route itself.  I really like the idea of breaking up the segment into a lot of much smaller fixed length segments and using this to generate a mean route sinuosity. Now to figure out how to make it happen....

DavidPike · Answer

Yes I did think the same was a simple effective solution, but a route could be indirect but still not very 'curvy'.  After a google, it did lead to an interesting tool which is the logical continuation of your idea Dan, which is to measure the sinuosity - but then do this with much smaller line segments (you'd need to split the polylines up prior of course).

Comparing the means of the 2 route's sinuosity indices would give an indication of relative 'curvyness' I guess 🙂

https://www.arcgis.com/home/item.html?id=00e708a448b74810a0e805c4a97f9d46

all really from this thread and linked threads:

https://gis.stackexchange.com/questions/195370/determining-curvature-of-polylines

https://gis.stackexchange.com/questions/37058/algorithm-to-find-out-points-of-inflection-for-a-polyline/37078#37078