I suggest adding a feature to “pause” data collection.
This might be useful during say street construction when the counts will be invalid.
In my case I’d like to take the device to another location when doing a manual count, and temporarily operate it there, without polluting the data at the home location.
Interesting idea.
In practice, roadworks would be something you would want to keep monitoring as you will potentially see a different mix, with more pedestrians and bikes. Keeping it active will ensure you collect the before/during/after data that will allow good comparisons to be made.
Moving the device is possible and if you do that and “install” it in that new location, then the data will be associated with the new street segment and not the ‘home’ one. That way it is effectively paused.
You can then restart in the home location when you are ready.
During roadworks a whole variety of equipment and people will mill about the area, without really being the same thing as the other data at the location.
Counting an adjacent street during roadworks is super interesting, showing the avoidance effect.
What I’ll do is take the device to my temporary manual count site,
record the current counts at the start, then calculate the delta.
This will however pollute the original location data.
I’m hesitant with a personal device to create a second account with a second email address to account for the temporary site, as that too will pollute the available data.
I need this feature too. We have a carnaval parade here (Vastelaovesoptoch). The road is closed, many huge objects pass slowly, several groups of pedestrians (music corpses) and lots of individual pedestrians everywhere, even in the middle of the road. I do not want the counting and recognition disturbed by this.
One could shut down the Raspberry Pi for a few hours, but this must be done properly by ssh. I can’t get this to work. Did post a separate question for this.
I can understand that this is not ‘typical’ traffic, but one of the points / benefits of continuous monitoring is that it can pick up unusual events like this and see the impact. I would not think that this will “pollute” the data as this will be a single blip on a day and not have a major effect on averages, or on the trends overall.
I want to ANNOTATE the data, publicly and in the API
10-Jan-2023 - Paving program starts
20-Jan-2023 - Paving complete
1-Feb-2023 - Roadway stripes done
10-Feb-2023 - Kidical Mass Bike Ride 5pm