Reassessing Traffic Accidents in Los Angeles - A Digital Humanities Project
Los Angeles Traffic Collision Data
The Los Angeles traffic accidents are collected by City of Los Angeles from 2010 on a monthly basis. This dataset covers traffic collisions in the city of Los Angeles from 2010 to the present (February 2022). It consists more than 500,000 datasets at the time of this project was made.
​
This project will feature 5 different pieces of data visualization. There were all produced using the Tableau application.
​
The second link provides the interactive visualization graphs from Tableau.
Accidents Over Time
The graph above shows the number of accidents over time starting from 2010 in various districts in L.A. There is a total of 21 districts based on the original dataset. Nevertheless, this graph shows a salient trend: the significant drop in traffic collisions starting from 2019. As anyone should have already guessed, the reason behind it might be the pandemic. The COVID-19 pandemic, which has infected several million people worldwide and killed hundreds of thousands, will be a disaster that the world will never forget in 2020. In order to prevent the spread of infection and potential deaths, many countries have implemented lockdowns. Although this had terrible economic and social consequences, one of the advantages of the COVID-19 control measures was a decrease in traffic flow on both urban and interurban routes, resulting in a significant reduction in traffic-related accidents. A similar effect had been found in several pieces of research conducted in Spain (Aloi et al., 2020; Nuñez et al., 2020), New Zealand (Christey, Amey, Campbell, & Smith, 2020), and Canada (Oguzoglu, 2020).
Map Distribution
The map distribution of the accidents has a similar content as the previous graph, but it shows the trend geographically. The map is produced by combining the original dataset with the LAPD division record file. The district was matched with the LAPD division accordingly. It might be difficult to see the shade of the color in the picture, it is recommended to inspect the map in the interactive workbook for a clearer view.
Age Distribution
The graph below shows the age distribution of the accident victims. It has a right-skewed distribution. The number peaks at around age 30 and decreases at a steady rate. However, one interesting feature from this graph is the victim number at the age of 99. The dataset does not explain this sudden spike. There are two possible explanations to make sense of this. One is that there is nothing wrong with this spike, and over 7000 of the victims in these accidents had an age of 99. Another explanation could be that the dataset only includes a two-digit number so that a victim with age over a hundred would be counted as 99.
Age Distribution with Sex
The graph below is plotted to emphasize the relationship between the age group of the victims and their sex. The figure shows that males in the 21-30 age group had the highest average traffic accidents (approximately 80,000 accidents). Data analysis also found that traffic accidents were more likely to happen to males than females in all age groups. Â Over the past 20 years, accident rates increased only among males in the age groups of 20-25 compared to all age groups and both sexes. The result of the 20-year data analysis also indicated the age group under 20 steadily declined in the number of accidents involved. In contrast, the 49-50 age group had the highest increase in the number of accidents in both genders.
Gender Differences in Number of Accidents
Besides the dataset from this project, several other articles have found gender differences in the accident rate. For example, according to the National Highway Traffic Safety Administration (NHTSA), men cause 6.1 million accidents per year in the United States, while women cause 4.4 million (NHTSA, 2022). Moreover, Insurance Information Institute (III) reported that male drivers were responsible for 37,477 fatal crashes in 2017, while female drivers were responsible for 13,502 fatal fatalities (III, 2020). These figures tend to support the claim that men are poorer drivers than women. However, a closer examination is needed.
Based on a statistic by Federal Highway Administration, 115.6 million women and 113 million men in the United States hold drivers' licenses (FHA, 2019). Moreover, men drive an average of 16,550 miles per year, while women drive an average of 10,142 miles per year, indicating that they drive significantly less than men (FHA, 2019). When these figures are added together, they show that women drive 30 percent less than men on an annual basis. Women had a slightly higher probability of being involved in accidents per mile driven than males, despite the fact that men cause more accidents.
Hence, there could be various reasons why men are more likely than women to cause accidents. First of all, men are more likely to be involved in accidents because they drive more kilometers each year. Furthermore, studies have found that men are also more likely than women to engage in risky driving behaviors such as driving under the influence of alcohol, not wearing seat belts, and disobeying traffic laws such as speed limits (Rhodes & Pivik, 2011). Moreover, the American Academy of Sleep Medicine Board of Directors (2015) had found that men are also more likely to drive while drowsy, with 56 percent of males have driven while drowsy, compared to 45 percent of women. Study by Harré, Field, & Kirkwood (1996) also found that men have a 3.1 percent higher likelihood of being penalized for reckless driving than women. For example, men were arrested for four out of every five DUIs prosecuted in the United States. Men between the ages of 21 and 34 are most affected. Despite accounting for only 11% of the adult population, this group is responsible for 32% of all DUIs in the United States. This could explain the traffic accident trend in the age group of 20-30 in the dataset.
Time Distribution Analysis
According to nearly ten years of traffic accident statistics study data, the distribution of road accidents within 24 hours of a day has an appropriate stability (Yu, Li, Chen, & Yi, 2009). Therefore, this project has also tried to assess the time distribution of the accidents and examine how it can fit with the previous research.
Based on the graph below, most accidents happened during the afternoon and early evening, which peaked at 1800 hours. This trend is consistent with the statistic from other datasets, where most accidents happen between 1600 to 2000 hours (Yu, Li, Chen, & Yi, 2009). Therefore, it is reasonable to isolate these periods as a high risk for traffic users, which allows transportation safety professionals to take appropriate actions to improve and increase the degree of safety based on the occurrence of each period.
