Which Hartford Corridors Have the Most Crashes?

by Isaac Frank and Chris Horkan

Last updated on 1 May 2023

for Data Visualization for All
with Prof. Jack Dougherty
Trinity College, Hartford CT, USA


Major and fatal car crashes occur frequently every day across the United States.

Speeding and reckless driving is not a new issue to Hartford: as the data in Ed Stannard's article finds that from January 2020 to July 2022, there were 614 crashes in 10 specified Hartford corridors, including 6 fatalities.

The article states reasons for crashes that are not unique to Hartford like, "human error, distracted drivers and non-motorists, driving under the influence of alcohol or drugs, speeding, and otherwise just not following the rules of the road (running stop signs, driving aggressively, not yielding to pedestrians or right-of way traffic, etc.)” (Stannard, Ed 4).

To Trinity students, this issue is important on a personal level because just last Spring a student died from reckless driving along a Hartford corridor.

This type of danger should not be in the minds of Trinity students or any one else in Hartford. This is why it is important to find the most dangerous Hartford corridors to eventually find solutions to remedy the danger they pose.

Given reasons that are not caused by infrastructure, we still located the most dangerous corridors within Hartford, and attempted to determine causality of crashes.

What is a corridor?

Simply put, a corridor is street, road, or avenue. In order to quantify the danger found in Hartford corridors, we had to view databases that contain Hartford crash data. The most helpful one we found was 2018-2022 Hartford police report crash data from the UCONN Crash Data Repository. Our data and subsequent visualization and interpretation is based on crashes taking place on their respective road, and other associated data like conditions and injury type.

It is worth noting that we did not include any non-local corridors such as interstates because interstate issues are not something that the city of Hartford has a large influence in, and would skew our data.

We also chose not to look at specific corridor intersections as it would likely skew our data and/or make crashes non-ascribable to a particular corridor. An analysis of intersection crashes in Hartford can be found here by Jess Cruz and Mia Rodriguez.

Crashes on Hartford Corridors and Normalizing by Length

So... which Hartford corridor has the most crashes?

It's Main Street.

As seen in Figure 1, Main Street in Hartford has the most crashes out of any corridor with 1,151.

Figure 1: This chart shows that Main Street has the highest number of crashes out of Hartford corridors with 1,151.

But is Main Street the most dangerous?

To answer that question we need to put our data into context.

You might expect a street called Main Street to be an important and long street. And in Hartford's case, you're right.

In order to adjust for this, we must normalize the data by corridor length (in miles). By normalizing the data, we mean to take the data and “adjust raw counts into relative rates, such as percentages or per capita," but in this case relative to length (Hands-On Data Visualization, Chapter 5: Normalize the Data).

As we build our model throughout this study, we must acknowledge that "It’s unrealistic to pretend that we can create a perfect model. But we can certainly come up with a good enough one" (Alberto Cairo, 2016).

Once we normalize our corridors for length, something is immediately apparent:

As shown in Figure 2, Park Street has the most crashes per mile.

Figure 2: This chart shows that Park Street has the highest number of crashes per mile with 368.

Types of Injury on Hartford Corridors

How many injuries (and what type) happen on Hartford Corridors?

As shown in Figure 3, the makeup of injury types in Hartford corridors look somewhat uniform, with some corridors leading others.

Figure 3: This chart shows injury makeup percentages caused from Hartford crashes.

The leading corridors of each type of injury are:

Possible: New Park Ave (67%; 32 total)

Suspected Minor: Farmington Ave and Granby Street (54%; 59 and 13 total, respectively)

Suspected Serious: Tower Ave (11%; 8 total)

Fatal: Tower Ave (2.7%; 2 total)

Main Street and Park Street, the leaders in total crashes and crashes per mile respectively, are not leaders in an injury type. Yet, Tower Ave leads both categories of Suspected Serious and Fatal Injuries.

With our insights, we can ask questions like, "is there something different about Tower Ave that explains why crashes reported there have higher rates of fatal or serious injuries?"

This is a question worth asking because although many crashes occur on longer streets, long streets may be much harder to prevent dangerous crashes on due to their seemingly "normal" injury correlation.

However, when there is an apparent anomaly of the serious injuries and deaths occuring on Tower Ave, we are in a position to find out why.

Causations and Limitations

Why is Tower Avenue so Dangerous?

Continuing on the Injury and Figure 3 discussion, we must also understand why Tower Avenue has such a high number of fatal and suspected serious injuries despite being the 15th longest corridor we looked at. This is because when you see an aerial view of Tower Avenue, you will see that the corridor is mostly straight with little public transportation, shops, and crosswalks. Compare this to other corridors like Main Street and Park Street, roads that are curvy with multiple public transportation stops, crosswalks, and shops, and we can understand why more people tend to speed and drive recklessly on Tower Avenue compared to other corridors. We can then use these aerial views and the data to find similar roads to Tower Avenue and identify other problematic roads. For reference, see Figures 4 and 5 below.

Figure 4: This map shows the difference between Tower Avenue, the blue corridor versus Main Street, the red corridor. This visual shows Main Street's curves compared to Tower Ave's relative straightness. Main Street also passes through downtown Hartford with many stop lights that reduce speeding as opposed to Tower Ave's propensity for speeding.

Figure 5: This street view of Tower Ave shows its lack of turns.

Methods & Limitations with the Data

While these insights are fantastic to find, there are many limitations with the data. It was mentioned earlier that the data used in Figure 1, Figure 2, and Figure 3 was taken from 2018-2022 Hartford police report crash data from the UCONN Crash Data Repository. It can not be stressed enough that police officers are reporting these crashes because police officers are not perfect and reporting these crashes is difficult. What makes a minor injury vs a serious injury is very subjective, which could skew the data we have avaliable to us.

From the data, police officers have to list over 30 observations, like Manner of Crash and Road Surface Condition, as well as the categories we mentioned above which is a multitude of observations to keep track of. Even though police officers are trained in the same way, it is very possible that not every accident was reported in the same manner and so the data and visualizations could be different if one police officer reported a crash over another.

This then leads us into our next limtitation, the way we collected this data. We collected our data in a similar manner to Jess Cruz and Mia Rodriguez who determined intersection crashes.

We first used our data from the UCONN Data Repository and downloaded it as a CSV file before uploading it into Mapshaper.org. From there, we used the console tab and input: -points x=Longitude y=Latitude. Then, oriented our map with the line: -proj wgs84 to geospatially fit our data into the correction positions. Each dot represented where a crash occured, as now we can get a better understanding of the corridors with the most crashes as a starting point.

From there, we drew corridors by hand in Placemark.io using a 20 meter buffer, and labeling each names of corridors using the name 'corridor' as a new property. We downloaded this as a .geojson file and loaded it into Mapshaper.org along with our point data from above. We then opened the console tab and wrote the line: -points x=Longitude y=Latitude once again to orient our data. We used the line: -join corridor-data fields='corridor' to join our data, and exported our new point data as a CSV.

The COVID-19 Pandemic

It is important to mention the effect that COVID had on the Hartford community. From the middle of March, about March 15, 2020 when government lockdowns were mandatory, there were 5,528 crashes reported in 2020, 5,448 crashes reported in 2021, and 5,294 crashes reported in 2022. This is a sharp decline from the 6,178 crashes reported in 2019.

Keen readers might notice that 2018 was ommited from this observation and that was a result of the limitations of our data. From our data, there were only 783 crashes reported in 2018 which does not make sense. This is too signifigantly low to include in the above finding.

Since more people were staying inside because of the government enforced lockdowns, it should come as no surprise that there were less crashes reported. However, this does skew our data as we will never know if the number of crashes had decreased because of increased safety measures by the city planners or because the lockdowns forced people to drive when it was only necessary.

Concluding Thoughts

From the data and visualizations, we can determine two key observations. One is that the longest roads typically have the most crashes, even when normalized for number of crashes by mile. We saw from above that Main Street leads all Hartford corridors with 1,151 crashes from 2018-2022 and Park Street has the most amount of crashes per mile with 368 in the same time period. We must also note that Park Street has the second most crashes and Main Street has the second most crashes per mile which is interesting considering that Main Street and Park Street are the two longest corridors in Hartford.

The second is that roads that have more turns and public safety measures have less crashes that lead to serious injuries or fatalities. Another way to put it is that roads with more serious injuries and fatalities, like Tower Avenue and Fraklin Avenue, are not only more straightforward, but also possess less natural stops which lends itself towards speeding and reckless driving.

However this is not a simple problem to solve. Multiple factors are at play when trying to solve the same issue that is discussed in the article "Connecticut’s roads are deadlier than ever. Figuring out why is complicated." by Katy Golvala and Dave Altimari. The authors list many potential reasons for these fatal roadways including, "People are driving faster", "Impaired driving is also playing an increased role in fatal accidents", "... there is also a mental health response to the pandemic at play on roadways", and, "Police are also giving out far fewer tickets" (Altimari, Katy Golvala, Dave) which means the root of the problem can not be pinned on one thing over another and prioritizing these issues is difficult for the jobs of city planners and law enforcement. The article goes more in depth about the reasons before, but the point is that this issue is complex and there are a number of factors at play for Hartford's dangerous roadways.

Works Cited

Altimari, Katy Golvala, Dave. “Roads Are Deadlier than Ever. Figuring out Why Is Complicated.” CT Mirror, 6 Feb. 2022, http://ctmirror.org/2022/02/06/connecticut-traffic-deaths-fatalities-dui/.

Connecticut Crash Data Repository. https://www.ctcrash.uconn.edu. Accessed 18 Apr. 2023.

Dougherty, Jack, and Ilya Ilyankou. Normalize Your Data | Hands-On Data Visualization. handsondataviz.org, https://handsondataviz.org/normalize.html. Accessed 18 Apr. 2023.

Stannard, Ed. “Here Are 10 Hartford Roads with the Most Crashes since 2020. With a Map That Shows Worst Spots.” Hartford Courant, 18 July 2022, https://www.courant.com/2022/07/18/here-are-10-hartford-roads-with-the-most-crashes-since-2020-with-a-map-that-shows-worst-spots/.