Public Transit Ridership
Source: American Public Transportation Association (APTA) (private nonprofit organization).
The TSI uses monthly American Public Transportation Association (APTA) data for the transit component of TSI for years prior to 2010, and data from FTA (Federal Transit Administration)’s NTD (National Transit Database) for 2010 and beyond. FTA is an agency of the United States Department of Transportation. The number of unlinked passenger trips is the measure used for the TSI.
Transit modes, include, among others, bus, trolleybus, vanpool, jitney, and demand response service; and heavy rail transit, light rail transit, commuter rail (including Amtrak contract commuter service), automated guideway transit, inclined plane, cable car, monorail, aerial tramway, and ferryboat.
Monthly data is reported to NTD by transit agencies.
The U.S. Federal Transit Administration requires that annual unlinked passenger trips and passenger-miles data be collected or estimated by the predominantly large and medium-size transit agencies participating in its National Transit Database (NTD). Monthly data collection was introduced as a pilot program in 2002. Over time, most transit properties developed new internal data collection and processing methods to meet the new requirements. These developments, combined with the implementation of more sophisticated validation checks by FTA, have resulted in more complete and accurate data in more recent years.
All ridership data reported relate to trips, not to people, because that is how data are collected and reported. The use of passes, transfers, joint tickets, and cash by people transferring from one vehicle to another, one transit mode to another, and from one public transit agency to another makes it difficult to count people. Boardings (unlinked passenger trips) can be counted more accurately. At the largest public transit agencies, even boardings may be estimated for portions of the ridership.
The majority of people using public transportation take two trips per day (one to a destination in the morning and one home in late afternoon or evening). A small portion, perhaps 5%, makes only one trip. Larger portions take more than four trips per day. At most agencies (10% to 30%) of riders transfer to a second transit vehicle to reach their final destination.
Data Quality Questions
After the close of a month, transit properties have one month to compile and submit data to the NTD. Therefore, there is a 30 days lag from the end of the month to the time the data is submitted. Some data may be missing for the most recent month if a transit agency failed to report data on time. Once the data is submitted, it is subjected to review and analysis for data completeness and reasonableness. Additionally, transit properties may revise their data at any time during the calendar year reporting cycle, which lasts through March 1 of the subsequent year. These changes may be done unilaterally by the transit property, as the transit property collects additional data on its operations, and these changes will be reflected in subsequent release of the monthly data base.
FTA requires properties to report a 100 percent count of unlinked passenger trips if the data is available and reliable. However, there are many properties that do not have the technology or the means required to perform a 100 percent count. (This is because the reporting of unlinked passenger trips requires an accurate count of the number of transfers made by passengers.) In other cases, the service is too complex to allow a 100 percent count. In these cases, FTA allows reporting based on sampling. In general, FTA does not know what method properties use to determine unlinked trips, i.e., if it is based on sampling or a 100 percent count.
- Is this data source a frame or sample?
- Does the sample cover the entire frame? Or is there some group that is missing or underrepresented in the sample? For the excluded group, what percentage of the total do they make up?
- Are the data available monthly?
- How soon the data are available after the month is over?
After the close of a month, transit properties have one month to compile and submit data to the NTD. Therefore, there is a 30 days lag from the end of the month to the time the data is submitted. Some data may be missing for the most recent month if a transit agency failed to report data on time.
- Are the data easy to access and use?
- Are the microdata available for use?
- Are there duplicate records?
No. Once the data is submitted, it is subjected to review and analysis for data completeness and reasonableness.
- Are there outliers in the data?
- Are data missing for individual records? If so, how are they identified?
- How accurate are the key fields (i.e., miles or trips)?
The key field is unlinked passenger trips. Accuracy depends on the individual agency and the data collection system used.
- Are variances available for this data source? If so, what method was used to calculate variances?
- Are the data comparable over time within the data source? If not, can data be made to be comparable (i.e., combining two data series)?
The date for the years before 2010 is from APTA. The data for 2010 and after is from the NTD. The two sources are comparable.
Other Questions and Important Information
- Is sufficient documentation available for the data source?
Data on the NTD can be be found at http://www.ntdprogram.gov/ntdprogram/data.htm.
- Estimation methods:
For small agencies, it is easy to report the actual ridership, for some large agencies, the number of boardings would be estimated for a portion of the ridership.
- Are other sources available for the same data?
Monthly data are available from APTA.
- Who is the contact for the data source? John Giorgis at FTA, 202-366-5430