Is the proportion of crashes that occur on weekdays greater than weekends?
By Mahdi A & Mahdi Y
Statement of the problem
We are conducting this project to see if there is a relationship between the days of the week compared to the days people get into car accidents. We’re interested in this because we feel that people drive differently on certain days. For example people may drive more carefully and less crazy on a Monday where people are just trying to get to work or school and back, whereas on a Friday or Saturday, people may get more energetic and reckless and are trying to have more fun and might get into an accident. With this study we intend to show everyone if there truly is a relationship between the days of the week compared to the days of car accidents.
Abstract
We choose to do a study based on whether or not it’s more common for people to get into a car accident on a weekday rather than a weekend. We chose to do this to see if it is true and that there is a relationship between the 2 times of events. From the population of Monroe County between the years of 2003-2015, we randomly selected a sample size of 5,000 accidents. We searched for datasets that have the day and year of the accidents but kept getting useless information until we finally found the eight right with correct information. We categorized this data by the weekends and the weekdays. We took our data and ran multiple tests and made graphs and equations.
Background Research
During the entirety of this project we went through many topics before we found the one that we liked. Many Data sets that we found before we landed on our current one were not useable. The reason for some of those data sets being unusable would be that the data they gave was summarized. After many days of searching for a good data set we finally found one when we searched up the term “Car Crashes ” on the website Kaggle. After we searched that we looked through multiple data sets until we found one that we liked. The data set that we landed on was labeled “CarCrashDataset”.
The data was taken from real accidents over the years of 2009- 2015 that happened in Monroe county Indiana. The data set did not provide a state of where it took place but it did include latitude and longitude coordinates which helped me locate where these crashes were occurring. This data set offered over fifty thousand samples (53,943 to be exact) and was very thorough with all of the different categories it gave. Some of those categories were the year, month, day, and even hour of the crash. The data set also included values such as the latitude and longitude of the crash, how many cars were involved and even the injury type that happened.
After we found our data set we then took a sample of five thousand from it to conduct our research. The data set included numbers that correspond with the days of the week. Once we saw that it included numbers we decided to see whether or not car crashes occurred more on the weekend rather than the weekday. While looking through the data though we found a problem which we had to fix. In our data there were some imperfections where a number that is supposed to correspond with the weekend is showing up as corresponding with a weekday. For example the number 6 is supposed to correspond with a weekday, but there were cells in the data set that said it corresponded with a weekend. Once we identified this problem we went in and fixed all of the imperfections without interfering with the data.
Ho: The proportion of car crashes that occur on the weekend is greater than the proportion of crashes that occur on weekdays.
Ha: The proportion of car crashes that occur on weekends is not greater than the proportion of car crashes that occur on weekdays.
The way that we will test our hypotheses is by using Minitab and using the 1-proportion test and compare the proportion of crashes on the weekend to the proportion of crashes on the weekday.
Proportion Of weekday Crashes
Proportion of Weekend Crashes
Graphs
Our Conclusion and Discussions.
We failed to reject the hypothesis that the 0.05 of significance is greater than the alpha Since the P-value is The proportion of crashes that occur on the weekdays is significantly greater than the proportion of crashes that occur on the weekend. Therefore, we have found sufficient evidence that there are more accidents on weekdays than weekends. The percentage of crashes that happened from the random sampled we took was greater than 50%, more than 70% in fact. We suggest more research and studies for a more accurate conclusion.
Minitab Data
This is a sample of our sample since our sample was 5000 values