Activity data is great! Very informative. There are all kinds of possibilities of parsing behavioral modes and getting into the nitty-gritty of diel behaviors from activity data. I would start with heavy exploration, e.g. visualizing against time-of-day, before and around important events, etc. You can look simulataneously at X, Y and sqrt(X^-37 + Y^2) - a kind of net activity index.
A possible methodological contribution would be to validate some coarse movement inference with activity data, e.g. to see whether the "encamped" states are much lower activity - they may or may not be at all (if, e.g., being "emcamped" includes moving around for other reasons). In any case, I would suggest just rolling up your sleeves and digging right in. I attach an analysis of panda bear activity collars that I did when Andy and I visited China - there it was to see if we can identify pregnancy and giving birth, but it gives you an idea of some approaches (and basic code).
Activity data is sampled every 3 minutes in raw data and once plotted against time, there are a lot of variations. So in the following figure, I have computed the net mean diel activity for C1 and L . By net, I mean the RSS (root sum of squares) of activity along x and y axes. I suppose the accurate unit for activity is degree as the activity along x and y axes range between 0 and 360. The figure clearly shows that C1 is much more active than L which means it moves around its head and thus its collar's accelerometer more frequently. The linear regression of net activity against time for both C1 and L is significant. For C1, I clearly see a trend and the regression line is steeper which indicates as it approaches summer (June-July-Aug) and the temperature increases it becomes more active. I also see a sinusoidal periodic pattern in C1's local regression graph especially between Mar-May but I don't know how to interpret it!
Parsing behavioral modes shows that while the animals are moving, the net activity values from the sensor are higher and vice versa. This means when the animals are encamped in a limited spatial range they move less (or they have fewer body movements) and vice versa.