Roller Coaster Blog Post

Introduction

Roller coasters are an exciting pastime and summer activity for many. While some people have limits on what they are willing to ride, some actively search for the biggest and baddest coasters for their next thrill. Just like everything else, roller coasters have appeared to evolve over time. In this blog post, I’m going to try to see if I can identify any trends in the engineering of roller coasters based on building material, type of coaster, and manufacturer of coaster.

The data set I am using comes from Reddit. The creator is Kyle Broxton. I appreciate him allowing me to use his data set. In this set, there are 1270 total roller coasters from around the world. 1005 of these coasters are currently operating, 224 are coasters that have been closed, and the remaining 41 rows contain data on coasters that are still under construction and unopened.

Each coaster has a rating and rank based off of how many of the other coasters on the list it is ‘better’ than according to 22 public online sources. Each coaster also has data for things like its top speed, max height, tallest drop, drop angle, and duration, as well as classifiers about what park its in, material its constructed from, type of coaster, and it’s manufacturer.

Trends

Building Materials

Before looking at any data, the most obvious likely trend that flashes out to me is what roller coasters are built out of. You may have a preference between wood or steel coasters, but I would hypothesize that less wooden roller coasters are being built and more steel roller coasters are taking their place. Steel coasters are likely more durable and parks can count on them to be longer term investments. It is also easier to create taller and faster steel coasters than it is for the same wooden counterpart.

As you can see in the graph, many more steel coasters started to open compared to wooden coasters around the 1970s. Out of the 41 roller coasters in the data set that have not opened yet, 40 are steel coasters and 1 is wooden. This aligns with our expectations.

While I believe that this data accurately shows the increase in the amount of steel roller coasters over the years, I think that it underestimates the decline in wooden roller coasters being built. The data set used for this blog is not a random sample of the population of all roller coasters. It is much more likely that the worlds most popular, the biggest, and the fastest coasters, or coasters that are notable for other reasons (potentially bad reasons) are in this data set than smaller and very average coasters. Because of this, you can’t assume that any trend I find in this data is representative of the population of roller coasters. However, it is representative of the sample used.

Before moving on, I will take a quick look to see if I am correct to think that steel roller coasters can more easily be built faster and taller than wooden roller coasters can be.

From this graph, we can tell that steel coasters have been built faster and taller than wooden coasters, which is an unsurprising result. We can also see that the max height and the max speed of a roller coaster appear to be highly correlated. This intuitive finding can be useful if we run a regression using either variable, as we know we should control for the other.

Type

I will now look into the different types of roller coasters.

Most Common Coaster Types
Coaster Type	#	Coaster Type	#
Sitting	49	Launch	39
Wild Mouse	48	Enclosed	37
Looping	47	Hyper	36
Spinning	42	Boomerang	26
Inverted	39	Euro-Fighter	24

There are 117 total coaster types listed in the data set. Excluding wooden, junior, and family roller coasters, these are the 10 most common type of roller coasters in the data set. Sitting, Looping, and Hyper coasters are all much more general types of roller coasters. The coasters within these groups show much more variability than the other groups on this list, which are much more specific. For example, the Interquartile Range for the max drop of Sitting type coasters is 70 feet, while the same measure on Wild Mouse type coasters is just 11 feet. The large differences in IQR still appear between these coaster types when you compare other variables, such as speed or duration. I am going to exclude Sitting, Looping, and Hyper coasters from my type analysis because of the larger variability in their data, and instead focus on the other 7 types listed. I am also going to include Accelerator and Bobsled coaster types in my data when looking for trends. They are both very specific coaster types, and are in the top 15 of occurrences in the data set.

An example of a Wild Mouse roller coaster type. These coasters are characterized by their tight unbanked turns and cars designed to be wider than the track. Coaster to coaster, Wild Mouse coasters are very similar.

Lets first look and see how often different coaster types were built over time. After plotting all 9 types of coasters, I narrowed it down to the 4 that were most interesting to improve the readability of the graph.

The first thing that jumps out at you when looking at all 9 coaster types, there is a heavy pickup in operating coasters between 1990 and 2000. This is likely a selection bias of coasters in our data set, as older coasters were probably less likely to be included, and this is the date range that the probability of being included increased.

Left to Right: Accelerator type coasters are categorized by their long straight launch runway at the start of the ride. Enclosed types coasters are built inside of a structure made soley for the ride, but are not the same as indoor coasters. Euro-Fighter type coasters have a main characteristic of a beyond vertical drop.

Another interesting takeaway is that Wild Mouse and Enclosed coasters have been around much longer than Euro-Fighter and Accelerator coasters. We can infer that this is because both Euro-Fighter coasters and Accelerator coasters both rely on relativity new technology that was not around when Enclosed and Wild Mouse coasters hit the scene. When trying to predict trends on what types of roller coasters will be built in the future, we should realize that the technology may not be invented yet for the most popular coaster types 25 years from now.

When comparing Wild Mouse and Enclosed coasters, which have been around for roughly the same amount of time, it appears that Wild Mouse coasters were a much stronger trend in the industry, and many appeared at once. In the past ten years though, it seems that very few new Wild Mouse coasters are being built. Enclosed coasters on the other hand, have been steadily increasing in number for roughly 40 years. I suspect that new enclosed coasters are being built because they can better adapt to new roller coaster technology to improve, while Wild Mouse coasters are what they are, and new technology doesn’t make them much better. Therefore, there is no reason for a park to build a ‘bigger and better’ Wild Mouse coaster, because they wouldn’t be that much bigger or better. It is also possible that consumer taste has faded away from Wild Mouse coasters, and for that reason parks have built less of them.

We can also compare the newer Accelerator and Euro-Fighter coasters to each other. They both appear to come on the scene around the same year. In fact, Accelerator coasters are built more than Euro-Fighters to start out. However, The number of operating Accelerator coasters flattens out quickly, and even begins to decline, while the number of Euro-Fighter coasters continues to rise. After looking into both types of coasters, I hypothesize the reason that one trend continues and one faded off quickly is the reliability of each type of ride. Accelerator coaster types are often considered less reliable than other types of coasters, and tend to require higher levels of maintenance. This could explain the flattening of number of operating Accelerator coasters, where parks decided it was not worth investing the money in one. If this is true, we can believe that reliability and maintenance cost have a large impact on the type of roller coasters ownership decides to put in their park.

Manufacturer

I will now transition into searching for trends based on Manufacturer. I will start by seeing if there appears to be a relationship between the Roller Coasters rating and its manufacturer to see if there is a class system in coaster making, and if some companies produce better products than others.

Each dot in this chart is a roller coaster rated in the top 50 of roller coasters worldwide. They are grouped by manufacturer.

As you can see, of the top 50 rated roller coasters in the world, the majority of them were manufactured by either Intamin or RMC. I want to focus in on those two manufacturers and see if they produce high quality rides exclusively, or if they produce tons of rides in general, so the are bound to wind up with some near the top.

From this, we can infer that RMC is a coaster firm that produces top tier products, and Intamin is likely very reliable and high quality, which would explain their very high demand.

From this point on in this section, I am going to be looking for trends using just the manufacturers listed in the table below, which is the Manufacturers with the 10 most coasters produced.

Most Common Coaster Manufacturers
Manufacturer	#	Manufacturer	#
Vekoma	184	Gerstlauer	69
Intamin	133	Schwarzkopf	65
B&M	126	Maurer Rides	36
Mack	91	S&S	34
Arrow Dynamics	83	Zamperla	34

I want to see if certain manufactures specialize on certain aspects of ride making, so I’m going to plot manufacturers against speed this time.

This graph shows us that between manufacturers, nobody tends to be dramatically materially different from the others when it comes to the stats of their roller coasters. This plot is very similar to the plot produced when manufacturers were compared by height, drop, duration, length, and drop angle. We can interpret from this that there are likely few trade secrets in roller coaster manufacturing that can allow your company to gain an edge in a particular area. It appears that if you develop technology to allow you to produce faster roller coasters, it will be adopted by other companies, and you will not maintain your advantage.

Conclusion

When looking for trends in roller coaster engineering, you can take a few things away from this analysis. The first is that wood coasters are being replaced with faster and taller steel coasters. This is an obvious sign of advancement in the field, so if you love wooden roller coasters, savor them while they’re still around. Second, the types of coasters that parks tend to build are likely at least partially determined by the current trends in technology, and their reliability as a long term investment. Lastly, and somewhat surprisingly to me, roller coaster manufacturers don’t seem to specialize in any particular area. If you assumed that one manufacturer may have the market cornered on building fast or tall roller coasters, or coasters of a particular type, you’d be wrong.