What common step metrics should sftrack calculate.

[birderboone]

The step_metrics() attempts to calculate common metrics between successive steps in an sftrack/sftraj. Currently those are:

• dx
• dy
• distance
• speed
• absolute angle

What other standardized metrics should be calculated?

[birderboone]

From a conversation from @rociojoo . Relative Angle should be added as well.

[pratikunterwegs]

Hi, just jumping in here, I wonder whether dt (difftime) is something that's already present and if you might want to add that? Might help users separate data based on large time intervals, which could indicate missing data in between.

[rociojoo]

Yes, dt is already there :)

[basille]

• Also to be computed, turn angles. Also known as "relative angles" (adehabitatLT) or "turning angles" (Turchin for instance). (I personally advocate for "turn angles"). Assumes ordering of steps. If an sftraj object is provided, there needs to be a check that the current step's start point is the same as the previous step's end point — but I would suggest not to reorder (can throw a warning if steps are not in order).
• Each metric should have their own vectorized function, such as step_dx, step_turn_angle, etc. as to only compute one metric on the file as a vector.
• As step_metrcis returns a data.frame, it would be good to be able to select which metrics to return — and possibly leave the possibility for the user to rename those in the arguments.

[basille]

Closing this issue now that sftrack is developed and on CRAN.