Problem:
We currently have no way to systematically recognize when an agency's number of trips is varying wildly (which normally suggests faulty-data)
Solution:
This method recognizes when an agency's number of trips on a given day has changed significantly from expected values for that day of the week (or for a holiday). If the new value is not within 10% of the normal value for that day, and is not on an exempt list, a message is sent.
In addition:
Each day has the number of trips taken logged
There is a suggested update to the current expected values for each day (that is set to approach the average number of trips on a given day when run each hour over the course of 3-5 months). In future iterations, the plan is to have this suggested update be written to the file from which it is read as part of this transformation.
Problem: We currently have no way to systematically recognize when an agency's number of trips is varying wildly (which normally suggests faulty-data)
Solution: This method recognizes when an agency's number of trips on a given day has changed significantly from expected values for that day of the week (or for a holiday). If the new value is not within 10% of the normal value for that day, and is not on an exempt list, a message is sent. In addition: Each day has the number of trips taken logged There is a suggested update to the current expected values for each day (that is set to approach the average number of trips on a given day when run each hour over the course of 3-5 months). In future iterations, the plan is to have this suggested update be written to the file from which it is read as part of this transformation.