I have a browser-based app I use where students wiggle the mouse and see graphs of position, velocity, and acceleration. These are the tasks I use with this.
Your goal is to produce an x-t graph that looks like a staircase going down and to the right. Discuss with your group (a) how you would need to move the mouse in order to accomplish this, and (b) how the v-t graph would look.
Pass out blank graphs, one per group, and do laps around room.
As each group agrees on a prediction, tell them to type in the URL they see on the demo screen and try it.
Imagine -- but don't do it yet -- that you repeatedly wiggle the mouse up and down, doing it fast but smoothly. (You will actually find it physically smoother to move the mouse rapidly in a circle; the horizontal part of the motion will be ignored.)
As a group, predict what the x-t, v-t, and a-t graphs would look like.
It takes some physical practice to get good results, e.g., for task 2 they often do it too slowly at first. The quality of the results is limited by the poor resolution of the mouse.
There are also sonar sensors that can be used for this. They're expensive and have their own difficulties and limitations.