Visualizations Using the UTA Data​
Here we have a basic visualization of the transit network in the SLC valley excluding the light rail and Forerunner networks. Each line is route is colored uniquely and its thickness is determined by the boardings. The stops are also placed along the lines and colored uniquely, although they are not colored according to the route. They are also sized based on boardings (diamonds) and alightings (pentagons).Â
Here we have boardings totaled across the whole year from 2017 to 2022. We get a great visualization of ridership growth across time and space. We can also very clearly see the impact of Covid during and after 2020. The dataset actually allows us to differentiate down to the month, but for the sake of simplicity I have left it to just the years.
Using just this visual data we can do some pretty basic analysis and say that downtown is definitely well connected and well used. The primary bus routes appear to be Redwood Rd, 3500S, State Street, and 900E. The primary stops are at the University and Front Runner and Light Rail lines. There is one thing about this map that can be a bit misleading. When looking to compare the size of lines, longer lines with more stops may seem at first stronger. However, we should also take into account the number of buses serving each line.Â
At first it would seem that lines like South Temple are much stronger than 5400S. But South Temple uses a total of 8 buses to move 2071 riders while 5400S only uses 3 buses to move 906 riders. Thus, making the riders/bus ratio for South Temple vs 5400S quite similar at 259-302 respectively. In this case 5400S is actually quite a bit stronger using this metric. The primary thing that I want to show when highlighting bus to rider ratios is the actual cost of a ride on a route.
The following maps compare average weekday boardings during April 2023 to boardings per bus. The first map shows all lines whereas the second only shows static bus lines that don’t allow for frequency changes or route deviations.
State Street Bus Lane​
When looking at both the map of the bus ridership and ridership per bus we can see that there are around 4 major routes. 3500S, Redwood Rd, State St, and 900E. Each of these routes serves 15 min frequency and connects directly to light rail stations. Specifically let’s focus on State St. State St uses 10 buses and runs from the Murray Central Station to the North Central Station primarily along State St as well as circling the Capitol. A one way trip from either end would take, according to the schedule, around 50 minutes while riding the bus. Using Google maps we can get some estimates for travel time using different modes at different times. Here’s a table showing those comparisons.
In all cases it is much quicker to drive than to take the bus. State Street does not have any dedicated bus lanes so at best we can expect the bus to move as fast as a regular traffic. For us, the worst case would see a ~20 minute delay. Just for the 8am Murray to NT we would see a 21 minute increase equating to a 50% increase in time spent. Now obviously if you were actually travelling to and from these two stations you would just take FrontRunner or the Blue Line but this does give us an idea of what someone farther offset from these two stations but still wanting to travel up and down State St. could expect in mode speed. We can use car traffic at 4am as sort of a baseline no traffic case. This would be a state similar to what a bus might expect in a dedicated bus lane. We should also include the time needed to stop and pickup passengers. This time should be pretty much constant and we will assume it is equivalent to the time difference between the car and bus during peak hours. During this time, traffic, not the schedule, should be the bottleneck. Let’s focus on 8am since this is likely when time is the most important.
Okay that’s a pretty impressive change in travel time. That puts riding the bus as actually being quicker than driving. That puts the new frequency at around 10 minutes and would be equivalent to adding 3 more buses.
All very nice in our imagination but would this actually be possible? Well State St. has very large sections of free curbside parking. The side of the curb is about 11 feet which is just barely enough to fit a bus. Certainly a bit tight, and lane adjustments would likely need to be made for long term service. Still that much space and the fact that free street parking extends for the vast majority of the route means there is a some serious opportunity for a pilot project. Now I’m not going to stretch these shaky numbers to far, but let’s just make up some EXTREMELY preliminary quote for adding another line to the outside lanes and some basic bus lane markings. From some quick research I did, the cost of a line is about $0.07 per foot but that depends on types of “paint”. That puts two extra lines along the 9 miles around $7,000 and lets double that to include bus lanes marking. A total of $14,000? Cheap! Of course let’s all now take our mandatory spoonful of salt. All of this is really just to say, why not try for 6 months and just see what kind of ridership changes happen? Do we really need free parking more than what is essentially a BRT upgrade to the 3rd highest ridership bus line? I think not.
General Observations on City Modal Share
Most times when looking at modal share our goal is to reduce car use and increase usage of other modes of transit such as bikes, public transit, or walking. First let’s take a look at modal share in some large cities across the world and see what this looks like.
This table shows the modal share across 68 cities acording to wikipedia. Now lets see what kind of patterns we can draw comparing care ridership to other kinds of modalities across these cities.Â
Alright this is a pretty simple regression analysis across many cities which can’t tell us much about any individual city and how it was able to make changes in modal share over time. But it can give us some insight into where cities tend to end up after building a transit network and how people get around within large cities.
However using this we can see that pretty much the strongest correlation with decreased car usage is increased public transit usage and following that increased walking. This is quite interesting to me as one of the primary focuses of recent transit activism is to increase bike usage to decrease car usage however this would suggest that it is very possible to have a low car usage without major bike usage.
Heres another table and set of graphs using data from smaller cities between 1 million and 250 thousand.Â
And the same regression analysis with this dataset.
Here we see similar trends but a significant change in how bike modal share effects car usage. For these smaller cities it seems much more likely that cities with low car usage would also have higher bike usage. In total though we see much lower correlations for between all modes suggesting that smallers cities are much more diverse in how they can structure their transportation for varying modal share. Just eyeballing it though I did see that a lot of this data looks like it is filled with outliers and special cases. To check my suspiscions I ran a check to see which points would be considered outliers when running the regression vs car usage. Doing this I found these points to be outliers since they were 2 standard deviations from the regression line.Â
As we can see no single city saw an outlier in all modes. To be very fair I have chosen to show the car usage regression excluding both single outliers and just double outliers.
Some interesting findings from removing these outliers. First when removing just the double outliers the R for Cycling decreases. Usually when removing outliers correlation increases but when looking at the double outlier cities we can see that these are very heavy cycling cities. As we removed outliers we can see that the correlations become stronger (who’d of thought) but the primary thing I would want to focus on is that Public Transport gains a stronger correlation than all others and increases its strength faster. This would suggest that the most common alternative to cars is public transport not walking or cycling. Of course when making decisions about what any specific cities wanted to focus on they would need to decide that on their own and see what mode would fit their cityscape and citizens needs. Its not always best to use the most common solutions and when taking a look at our list of outliers we can see many very successful cities. Certainly Amsterdam while definitely being an outlier is a city many others and especially urbanists would like to emulate.
Using this data the strongest statement I will give is that to have a successful city which does not rely on cars you most often need strong public transport and walkability. Bikability is not required however when focused on can be successful as the primary form of transport.
UTA Transportation Tools Cost Analysis​
Let’s take look at the cost of different modalities UTA serves. UTA has roughly 5 different categories of transit it provides. Bus, Commuter, Light Rail, and “Other Transit”. Commuter is basically just Forerunner. Light Rail includes the Red, Green, and Blue Lines, as well as the S-Line. Bus is just basic static bus routes. Whereas “Other Transit” would include things like route deviation lines, paratransit, rideshare and vanpool, and other microtransit. I have detailed the total Operating Cost, not the total budget, for each of these categories. “Other Admin” includes any of the other general and administrative costs associated with operating UTA in general. Seeing as how a large portion of UTA’s budget comes from a Sales Tax which would, and has, increased along with inflation, I felt that having this comparison as well would be good to understand exactly how much people are actually paying. I have also detailed the cost per boarding. You should be aware that this is averaged across all lines and so some lines are dramatically more cost effective than others and visa versa. Thus, it would not be fair to look at any particular line and use this as a cost estimate. Alright with all those caveats out of the way here is the data, use it at your own peril.