Swarm robotics investigates the use of numerous, relatively inexpensive robots for spatially distributed tasks, such as foraging, exploration, and search and rescue. In a swarm, each robot only interacts with their local environment and immediate neighbors, giving rise to emergent behaviors that can be useful, but often difficult to predict. Furthermore, most swarm tasks involve a need for a human to correct, inform, and monitor the swarm during its operation. In order to facilitate such interaction in a simple waypoint following task, we monitor important state variables in a simulated swarm of 2000+ agents during operation--including algebraic connectivity and consensus about the user-given goal. These are used to create a time series model of the swarm with which to forecast the state in the near future. We can then discern what affect external interruptions (i.e. human input) have on the future swarm state, and use this information to better inform the human operators. Preliminary results will be presented.