Missing Trips and Service Metrics datasets

Swiftly has a couple of datasets that make it easier to understand missing service: Missing Trips and Service Metrics datasets. These can be found in the Download Center. We have included the Missing Trips dataset which covers many use cases related to missing service. We have also included the ability to add additional fields needed for NTD reporting.

Notes:

• Swiftly provides information about service that was observed and not observed (missing) based on location data received from the agency’s vehicles. We note data as missing when we do not have arrival or departure data for a scheduled stop. For customers using Swiftly’s Service Adjustments product to cancel service, this export will highlight service that was officially cancelled via a service adjustment (did not operate).  Note: This is related but different from providing data on whether or not service was operated.
• Swiftly provides tools for assisting with NTD reporting. This is different than providing rolled-up numbers that are ready to enter directly into the NTD system. That is because Swiftly should be paired with additional datasets to compute the most accurate data based on an agency’s specific methodologies for reporting.
• This dataset does not gracefully handle detours. If a vehicle is on detour and does not travel past stops outlined in the GTFS, those stops will be flagged as “missing” even if the vehicle was on detour. Additionally, geometry that strays from the static GTFS will not be included in captured mileage so this should be added if important to an agency’s calculations.

 

Missing service in Swiftly

The Missing Trips dataset includes the list of all scheduled trips with incomplete or missing stop data. This is a great list for a starting draft of “missed service”.  From here there are a number of columns that are useful for QA to go from “missing service” to “missed service” , ie taking the list of trips with missing data and verifying whether service was in fact operated or not.

1. Trip completion status - shows whether a trip has some or no data. Options are “MISSING” OR “PARTIAL”. A missing trip a trip that has 0 observed stops and any trip that has greater than zero and less than the  full number of observed stops is called partial.
2. Trip adjustments - this column gives insight into whether missing data is expected. For example if a trip has a known detour, it may have expected missed service, whereas if a trip does not have a relevant service adjustment, there may just be a data issue. Note: only agencies with Service Adjustments will see data here.
3. Number of scheduled/observed stops, number of missing unexplained stops - these columns can be helpful for digging in more to expected/ not expected missed service. The Num_missing_stops_unexplained field shows the number of stops that are missing and not associated with a service adjustment. This can help an agency figure out where they may need to cross-reference operational data to see whether there was a reason an adjustment was not detected or whether stops may be missing due to data issues rather than actually being missed.
4. Trip missing first/last stop - Similar to above, these can be useful for understanding why data may be missing. It is common for agencies to have some missing data at the beginning/end of a trip so this may indicate that trips were not necessarily missed and can be useful for identifying GTFS issues with stop locations or assignment issues that cause the trip to be unassigned prematurely, etc.

In addition to showing how many stops were served, this dataset shows the scheduled vs missed/not served running time and distance for each trip.

• Scheduled_runtime_seconds_not_served: sum of scheduled runtime for all stops without observed data. This shows the amount of scheduled runtime that was missed.
• Scheduled_running_meters_not_served: sum of scheduled meters for all stops without observed data. This shows the number of scheduled meters that were missed.

What information is included with the National Transit Database (NTD) metrics?

The "National Transit Database metrics" toggle includes all of the columns listed above in addition to a few additional ones that are useful specifically for calculating NTD metrics.

• Daytype: This is useful for understanding missed service by “weekday”, “saturday”, and “sunday”.
  • Note: We do our best to infer daytype from GTFS and will return all dayTypes that match a given serviceId. In cases where service is only defined in calendar_dates.txt we will return the dayType associated with the service date.
• Estimated scheduled layover: Is calculated by taking the minimum of the scheduled time until the next trip or 20% of scheduled running time. No layover is attributed for the last trip of a block which can be determined by seeing if there is a value for next_trip_in_block_scheduled_start_time.
• Estimated observed layover: Is calculated by taking the minimum of the observed time until the next trip or 20% of the scheduled running time. No layover is attributed for the last trip is a block which can be determined by seeing if there is a value for next_trip_in_block_scheduled_start_time.
• Scheduled revenue hours: scheduled_running_time + estimated_scheduled_layover
• Observed revenue hours: observed_running_time + estimated_observed_layover
• Scheduled revenue miles: scheduled_running_distance (m) /1609.
• Observed revenue miles: observed_running_distance (m) /1609

What information is included in the Service Metrics dataset?

This dataset includes a summary of service delivered and scheduled grouped by any number of properties including month/service_date/route etc. This dataset is useful for understanding the difference between service delivered and scheduled at the summary level. The key columns in this dataset are:

num_scheduled_trips: The number of scheduled trips in the group

pct_complete_trips: The percentage of trips in the group that are complete

num_missing_first_stop: The number of trips in the group that are missing their first stop

num_missing_last_stop: The number of trips in the group that are missing their last stop

num_missing_trips: The number of trips in the group that are completely missing (no stops tracked)

num_partial_trips: The number of trips in the group that are partial tracked (some stops but not all stops are tracked)

num_adjusted_trips: The number of trips in the group that have known service adjustments applied

total_scheduled_runtime_seconds: The total of the scheduled run-times of all the trips in the group

total_observed_runtime_seconds: The observed runtime of all of the trips in the group

total_scheduled_running_meters: The sum of the stop path lengths for all the stops on all of the trips in the group

total_observed_running_meters: The sum of the stop path lengths for all the observed stops on the trip. If a trip exits its GTFS-specified shape (ex: on a detour), that distance will not be included.

total_scheduled_running_meters_missing_trips: The sum of the stop path lengths for all the stops on all of the completely missing trips in the group. Total observed running meters for missing trips will be zero, since the trips were completely missed.

total_scheduled_runtime_seconds_missing_trips: The total of the scheduled run-times of all the trips in the group. Total observed runtime seconds for missing trips will be zero, since the trips were completely missed.