Visualization of 9 Years of Metrorail Ridership
Metro planning staff understand that a picture is worth at least a thousand words, and often more. (And that a video is worth 1000 * 30 words per second.) As such, we are always looking to increase our ability to create compelling graphs, charts and video simulations.
Metro planning staff adapted this code to create a visualization of Metrorail ridership data from 2004 to the present. (Link opens in new window.) Here’s how it works:
- Each year is a horizontal stripe, sectioned off into months that go across. Both years and months are labeled.
- The days within each month are transposed, so start at the left and read down, then move right. In the sample image, the leftmost column of January 2004 is the first week, with Thursday January 1 being the darkest red square. A graphical example is also displayed in the legend at the top of the visualization.
- Each day is colored according to the ridership on that day, with darkest red being the smallest range (0 to 99,999) and the darkest green being the highest range (greater than 1,000,000).
- If you move your mouse pointer over any individual day, a small “tool tip” appears showing the date and the ridership for that day, rounded to the nearest 1,000.
Looking at ridership data this way is very intriguing. First one notices the decreased ridership on weekends, and that Saturdays are usually higher than Sundays. Next, the red-colored weekdays pop out, indicating Monday holidays such as Presidents Day, Memorial Day, and then other holidays like Thanksgiving Day or Christmas Day. Once the logic of the display is figured out, the next impulse is to look for and explain anomalies. President Obama’s first Inauguration Day was our highest ridership day ever. Snowmageddon shut down the DC region for nearly a week. A green square the last weekend of October, 2010? The Rally to Restore Fear/Sanity. The higher-than-average ridership in April 2013 including a cluster of dark green squares representing mid-week days? A surge of visitors for the Cherry Blossoms over a two week period, including the 4th Highest Ridership Day on April 10, 2013, when 871,000 trips were made on Metrorail.
Zooming out from the micro to the macro, other patterns become noticeable: Higher ridership in June and July (in the 700K range) in 2004 starts to spread to other months over the next four years. Eventually, the base weekday ridership is in the 700K range. Then, June and July are regularly in the 800K range most days of the month starting in 2008 which starts to spread to other months, peaking in 2010. Starting in 2011, the darker green starts to retract, which continues into 2012.
So far in 2013, we seem to be on track to have ridership similar to last year. Where ridership goes for the rest of the year is up to a variety of factors.
What other trends are apparent to you in this visualization? What other useful visualization techniques have you seen implemented for transit data? What other Metro data would you like to see presented in this format? Leave your thoughts in the comments below.