Jitter and Latency is normally measured End2End and the source of the jitter and latency can come from multiple points across the path.
From the air interface jitter would in many cases be a result of either congestion and thus buffering leading to a bursty scheduling of the resources on the air interface for this user. The other alternative would be poor radio and excessive HARQ retransmissions giving you a jitter in multiples of the air interface one way latency.
Generally high latency (on all traffic flows) would be a subject to excessive buffering in the transport path (e.g. buffer bloat) for the RAN or Core transport. You should look at if you have certain congested segments and see if you can either upgrade the QoS settings of the VLANs used, upgrade link speed or move certain traffic to a different link. If you do not have visibility for where you have possible congestion you can try to run a set of TWAMP traffic streams (use the RBS or router as the reflector/end point) or maybe only do a continuous ping towards a relevant endpoint. Remember that much of the traffic in the network is bursty and thus you might see that the average (15min) load of the link is not that high but you can still have significant congestion due to microbursts.