Monitoring Passengers Loads on Metrorail – Using New Tools to Examine the Data

January 5th, 2016

The new version of the Line Load Application now models passengers into trains by cars. Let’s take a look at this new feature!

Remember in May when we said an updated version of the Line Load Application was coming that would include passenger distribution data at max load locations? Well it’s here now!

If you’ve seen Metro employees with clipboards out during rush hour at major stations, then chances are you’ve seen the Metro load checkers. These individuals mark down the loads of these trains. They also mark down any people who didn’t board. Last but not least, they are also doing this by car, and with that information Metro has been keeping track of the spread of the loads on the cars at the max load stations.


Average Car Loads in the AM Peak Hour – October 2014 Weekdays – Modeled Distribution of Passengers at Dupont Circle **The estimated railcar crowding is based on the scheduled Red Line service.

With the data from the Metro load checkers — passenger loads by car data –combined with origin-destination ridership data, the Line Load App assigns passengers to rail cars at the max load stations with the same distribution as the data collected by the checkers.  The updated version of the tool allows us to simulate railcar crowding conditions at a number of core stations, beyond the stations where traffic checkers collect data. For example, the graphic above shows the simulated distribution of average weekday peak hour (8:00 AM to 9:00 AM) passenger car loads for October 2014 at Dupont Circle.

This new feature can tell Metro planners about potential safety issues as part of station capacity studies. This could also be useful for future customer information and safety campaigns to encourage people to spread out further along the platform to board less crowded trains.

How else do you think Metro could use this data? What can we do to help spread the peak travelers to the less crowded railcars?

Related Posts:

  1. Peter K
    January 5th, 2016 at 10:01 | #1

    Any thoughts on why loading is higher on the longer inbound train? That’d make sense if the 8-car trains are more likely to be the ones coming all the way from Shady Grove, but otherwise it seems strange.

  2. Mr. Transit
    January 5th, 2016 at 14:04 | #2

    The big takeaway here is that the eastbound Red line service in the peak hour last October was well below the Board-adopted Loading Guidelines. I hope that Metro will provide this data, by line, at other peak load points.

    • January 6th, 2016 at 08:13 | #3

      Not quite, Steve. Remember that the figures above represent averages across an entire hour. With six minute head ways on Red, the actual passenger loads on individual trains range from extremely crowded (north of 120 persons per car) to “acceptable” (80-100 persons per car). The actual customers experience crowding conditions well above the loading standards, and if we could somehow estimate a weighted average, then these hourly averages would all skew much higher.

    • Melissa
      January 6th, 2016 at 12:43 | #4

      Another point to remember is that the LineLoad uses the schedule, so the results shown are, in this case, the average October Tu-Th loads placed into the scheduled trains, which might paint a rosier and less crowded picture than actual conditions. What’s not built into the app is actual train data for a specific day. We do, however, sometimes manually sub in that data so we can see something closer to actual conditions.

  3. Mr. Transit
    January 6th, 2016 at 11:00 | #5

    Well then post 15 minute segments to better inform. Variances from the scheduled headway will always produce occasional overcrowded and underutilized trains. That’s why schedulers build some margin into the train schedules. I also agree that showing the peak hour loading will mask a bit the most crowded trains and the least crowded trains.

    WMATA also needs to consider the amount of time that a customer experiences loading of over 100 passengers per car. If it is only from Dupont Circle to Farragut North or MetroCenter that may be fully acceptable. To schedule for only 90 to 100 passengers per car in the short downtown segment of the Red line would result in excess capacity at other points along the Red line and inefficient service delivery.

    Lastly, I correct yesterday’s post that referred to October, 2015; the information provided by WMATA is for October, 2014.

    • January 7th, 2016 at 09:47 | #6

      we’re getting there, steve. later this year we should have a new tool to show the relationship between actual train movements and actual passenger movements and give us a train-by-train (and platform by platform) assessment of passenger volumes, in real time. give me a few months to get this finalized but we’re close.

      the issue that you aren’t landing on is the relationship between an overcrowded short-distance train and the overall line throughput. in the core we are experiencing berthing times that are 25 to 35 seconds longer than scheduled. train operators can sometimes make up for this lost time during longer stretches (i.e. Rockville to shady), but that is encumbered by train pull outs and slots upstream. The net effect is that even short spurts of crowding-based dwell create line effects that hurt everyone. Red Line takes six minutes longer to run end to end than it did in 2011 – 67 minutes instead of 61 minutes. That’s a mighty capacity decrease – 10% – on the red line, in large part due to crowding-based dwell time increases.

      If you think it’s ok to inconvenience DC travelers for one stop, you’re not taking into account that EVERYONE pays the price in terms of the accumulated impact of increased dwell times on line capacity.

  4. Wendy
    January 6th, 2016 at 11:51 | #7

    Peter, the longer trains (8-car trains) are operated on the Red Line running between Shady Grove and Glenmont. The Red Line that goes from Grosvenor to Silver Spring only uses 6-car trains. The Shady Grove and Glenmont line has much higher ridership, carrying passenger who commute over longer distance to the downtown core. It makes sense to see a higher level loading on the longer trains.

    I ride Red Line at Gallery Place every day. In the evening, I can see the railcar crowding difference between the Shady Grove bound Red Line and the Grosvenor bound Red Line. The westbound passengers going beyond Grosvenor would not board the Grosvenor bound train.

  5. Ann W.
    April 9th, 2016 at 15:12 | #8

    Hi, is the Line Load App available for us to use and is it possible to talk to someone in detail about the algorithms behind it? I am working on my dissertation research and this is possibly applicable. Thank you!

Comments are closed.