This week, we are working with the Monthly Modal Time Series dataset to create a time series visualization using Tableau Public. After looking at the dataset, I decided to keep the 6 variables:
- Primary USA City
- Year
- Vehicle Revenue Hours
- Ridership
- Collisions with Motor Vehicle
- Collisions with Person.
I chose these variables because they allow me to compare collision trends over time and across different cities. I can look at changes over time, city comparisons, differences between collision types, and additional context about each transit system.
For my visualization, I decided to focus on collisions with motor vehicles and persons over time, based on city. To do this, after uploading my data in Tableau, I created a dual axis line chart. The columns/x-axis is the year, and the rows/y-axis is the number of vehicle and person collisions, and each is color coded. I also added a filter by city. This allows you to see the average collisions of all cities, a few, or just look at each one individually. Below is the interactive graph, and on the right hand side you can change the city being displayed.
Link here.
Looking at the graph across all cities, the first noticeable thing is that collisions with motor vehicles are generally higher than collisions with persons. This difference can be seen when comparing the two lines on the chart, with vehicle collisions usually having higher totals over time. The two types of collisions do not occur at the same frequency in the dataset.
The more interesting pattern is the spike in 2018, where both collision types rise sharply and vehicle collisions rise the most, and the sharp drop in 2019. The drop however, might be a reflection of incomplete data for that year rather than a real decline, so I would be cautious about reading it as a true change in collisions.
When I looked at the cities individually, a 2018 spike was actually not common. Most cities peaked in other years, some in 2016 or 2017, and others stayed fairly steady with no major spikes at all. The 2018 peak in the combined graph is driven mostly by a few large cities that did spike that year and pulled the total up.
New York, for example, records roughly 500 to 900 vehicle collisions and about 100 to nearly 300 person collisions per year, while a small city like Bend, Oregon records only 1 to 5 vehicle collisions and often 0 person collisions. Because the large cities have such high counts, they dominate the overall average graph, so the combined trend mostly reflects what happens in the biggest cities rather than a typical city. It is also important to remember that the chart shows collision counts, not rates. A city with more collisions is not necessarily less safe, since it may simply have more people, more vehicles, and more traffic. This is why looking at cities individually, or a few at a time, is more informative than relying on the overall average.
Leave a Reply