Violation Metrics

[zanebclark]

@jlarsonOmahaNE brought up a great point in #64.

we would want to be able to fine a vehicle for this action but also track how long they were in the no stopping area.

For both Session and Aggregate endpoints, we should support compliant metrics along with non-compliant metrics. Non-compliant metrics should include information about how much of the transaction was non-compliant.

It sounds like this would be useful for prioritizing enforcement resources and re-defining curb policies.

[jlarsonOmahaNE]

Are you thinking about adding something to session_type as a enforcement/violation tag? Or just use area and then add enforcement in metric_type?

[zanebclark]

I think there's room to combine this with #73.

If the park wasn't a violation, how long did the park have before it would have been considered one?
If the park was a violation, by how much?
What's the violation cost / penalty?

[jlarsonOmahaNE]

I think that if it wasn't a violation, it would depend on each city and what they would want to define as time.

For a violation, I think it would be some time past the max_stay like 1-5 min?

I think the violation cost would also depend on the rate each city has set. We issue a violation first in a tow zone then call the tow company right away and they have 30 min to get to the vehicle. We can only tow if it is signed Tow Zone. Other areas just get citations such as parking in a bike lane.

[zanebclark]

I think that if it wasn't a violation, it would depend on each city and what they would want to define as time.

Isn't Methodology supposed to be standardized across municipalities?

For a violation, I think it would be some time past the max_stay like 1-5 min?

It would compare the event duration to the policy. If the policy had a max_stay of 5 minutes and the vehicle parked for 20 minutes, the overage would be 15 minutes. If the policy had a max_stay of 30 minutes and the vehicle parked for 20 minutes, the overage would be -10 minutes.

[jlarsonOmahaNE]

Yes Methodology would standardize it across all cities. So would there be a new metric for enforcement dwell time? And would you need a new event_type to label enforcement? See current event types.

So tracking enforcement time would be something like: IF event duration > max_stay THEN sum[enforcement time] / count [enforcements]?

[zanebclark]

@jlarsonOmahaNE , I don't think we would need a new event_type. According to a note here:

A "session duration" or "dwell time" can be calculated by calculating the difference between the event_time_start and event_time_end.

I think we would to do two things:

1. Expand the definition of Session to include a reference to the policy, or policies, that were "used". Multiple policies couldn't be applied concurrently, but a park duration could extend across multiple policies. I'm not sure the best way to reflect this potentially one-to-many relationship in the Session object. For each applicable policy, there are a few metrics that might be useful:
   • time elapsed before policy was applicable
   • time elapsed when policy was applicable
   • time remaining before the park would have been a violation
   • time elapsed after the park was a violation
   • time elapsed after the policy was applicable
2. Introduce a new Methodology that reflects aggregates for these data points.