Metrorail Rider Incomes – A Closer Look

June 29th, 2015

Salaries of actual riders are needed to paint a true picture of Metrorail ridership by line.

The Washington Post recently featured a series of images from the You Are Here project of the Social Computing Group at MIT showing Metrorail median income by line and station.  We were digging into it and realized it uses median household income within a half-mile radius, and not that of the actual riders’ households.  While we’ve mapped low-income riders before, we set out to answer the question, “What is the actual average income of Metrorail riders by line and station?”   Along the way, we developed this interactive data visualization.


Screenshot of Metrorail rider income by station visualization. Click image for full interactive version.

The biggest overall difference between our work and that of the MIT group is higher  household incomes at end-of-line stations on the eastern side of the region.  These stations, while located in lower income areas, have large parking facilities that draw commuters from all over the region and beyond.

One additional feature of our visualization is the ability to toggle across different time periods of the day.  AM Peak shows the income of people entering the system in the morning, giving an idea of where they live.  PM Peak shows income of people entering in the afternoon, giving some insight to where the high paying jobs are located in the region.

Please note that the Silver Line isn’t shown, as the income data comes from the 2012 Rail Ridership Survey.

Can you compare this to the MIT project and find other differences?  Let us know what you find.  And if you’re feeling creative, download the Tableau workbook to access the data and see what other interesting visualizations you can make from it.


Related Posts:

  1. angie
    June 30th, 2015 at 15:17 | #1

    For clarification, this analysis does not use the same data as the MIT report, correct? I am confused by the title of the article which suggests that the data used in the analysis is not from actual riders.


  2. Michael
    June 30th, 2015 at 16:18 | #2

    Sorry if the post isn’t clear. The MIT work only used median household income for households located in a half-mile of the stations. Our analysis used self-reported income for riders by AM and PM entry stations from the 2012 Rail Ridership survey. I hope that clears it up.

  3. Douglas Stewart
    July 9th, 2015 at 14:25 | #3

    Thanks for this interesting data. I would have expected the Vienna – GMU data to have even more variation in income levels between morning, afternoon and evening riders. This station area needs to get more commercial and office development to stimulate more reverse commutes and make the Orange Line more efficient.

  4. Christine
    July 10th, 2015 at 02:24 | #4

    The difference here is MIT used the right metric for income, which is median income. Using average income is generally not considered an acceptable practice, as the median is considered to be more reflective of the economic circumstances of individuals. Whenever averages are used it is usually a red flag for me that the person does not have much of an understanding of how to do proper data analysis regarding the household income metric. This is what made MIT’s study significantly more useful, they used the commonly accepted metric when doing analysis regarding this type of economic analysis. Lesson here, never use the average for area income and mapping, use the median.

  5. Michael
    July 10th, 2015 at 09:58 | #5

    Thanks for the comment. Yes, we are aware of the difference between median and average incomes. However, the survey didn’t explicitly ask for a rider’s income, instead to tick a box specifying an income range. We had to interpolate an “average” value from all the responses. So in reality, this number is neither a true average nor a median but something else. However, we believe it’s much closer to the actual incomes of our riders than the MIT project’s output.

    Again, the difference between this data and that from the MIT project is we are using incomes of actual riders’ households, whereas the MIT work looked at income from households within a proximity of the station regardless of whether they rode Metro or not.

Comments are closed.