Hi @JoeD,
I used a camera in a very congested portion of our manufacturing production floor. I was trying to identify the times and specific areas where we had issues with line moves. I had to move a single camera several times to identify the best angle to get the most complete data. It worked well for us, but we also have quite a bit of head room in our building and I was able to get far enough away to get the complete spread I needed within the frame. I have moved the camera elsewhere since, but it worked out nicely. The Heat Map Analytics is still in Beta (I believe) and so I would think it should get a little better in the future. Currently, you can filter by hour or by day. I'll show you one of our cameras we have up now. Depending on how detailed the area of monitoring and total area coverage within the frame, I could see it being problematic for very detailed, specific areas of interest while using a single camera (especially if you have high traffic throughout the rest of the frame). One thing I hope they will add in the future is to be able to show a heat map of a video wall with multiple cameras. Other than the heat map itself, there is no true data analytics on movement that I can tell. Hope this helps.
One thing to take into account is that the heatmap colour range is not an absolute scale. It is relative to the total motion for that camera, that date, and the selected time range. What this means is if there is minor motion at 5AM, it may not show up at all if there is significant motion later on in the day and the time range covers a large period of time e.g. 4AM to 11PM. The motion at that time relative to the motion for the total time period is very low.
What you can do is try out different time ranges to discover small changes. In the example I have just given, choosing 1AM to 6AM will provide a much better chance of finding minor activity. This is because the activity at 5AM relative to the other low level of activity at night is more significant.
As has been pointed out, camera placement makes a huge difference in terms of the usefulness of this functionality. Cameras only understand a scene in two dimensions. The shallower the camera angle of view, the more occlusion will happen and it will impact the likelihood of the heat map correctly representing activity for an area. The steeper the angle the more effective the heat maps are at detecting a specific motion.
