For most people, going to an airport is already a dreaded burden that makes flying costly, uncomfortable, and awkward. However, despite these inconveniences, the ability to cross an entire country in a matter of hours is immensely beneficial for numerous reasons. Not only does it allow for people to quickly react to different situations, but it also opens up new opportunities for entrepreneurs, businessmen, engineers, doctors, and just about any other career. This benefit alone is what makes commercial aviation so successful in today’s world. Since time is valued so highly in today’s society, it is essential to constantly look for ways to become more efficient. However, despite the constant efforts to maintain a timely schedule by both airports and airlines, it is inevitable that some flights will be delayed, ultimately destroying the benefit promised by the aviation industry. Hearing that a flight is delayed 45 minutes or even canceled is not only disappointing to both the passengers and the airline, but also extremely costly. Understanding what can potentially cause these delays is essential to understand how to prevent them as well.
While they’re may be many factors that contribute to flight delays, one might expect that weather severity would create at least a moderate association to help predict flight delays. However, since the weather is always changing and flight delay information also changes by the hour, collecting this data manually is both inefficient and time-consuming, making the data irrelevant even after it is collected. Therefore, to create models that are always current, it is essential to collect data with a source that can quickly go through lines of information in a matter of minutes rather than a matter of hours.
This can easily be done with a series of API files and a computer programming language such as Python. API files are very useful to computer programmers as they provide numerous amounts of data, both old and current, that can be used by programmers to make statistical models or allow them to understand the current situation of a certain topic. In essence, API files are very useful to computer programmers as they allow one program to access data from another source. There are hundreds of API files available to the public on the internet, each pertaining to a unique subject, such as countries, prices, quality of life, and so much more. For this project, the API files that would be most relevant would be those containing data that is not only updated and current about flight delays at multiple major airports in the United States but updated and current information about weather information at those major airports as well. By plotting the current weather severity ranking, which is just the cumulative score of visibility, wind, precipitation, freezing, and danger at the airport, each of which are given a score of 0-10 allowing for a maximum score of 50, at each of the airports on the x-axis, and then plotting the flight delays at those respective airports on the y-axis, one would be able to create a scatterplot of the data and create a Least Squares Regression Line to help see if there is any association between flight delays and weather Severity. After creating a program that would allow the user to go through both API files and graph the data (https://trinket.io/python3/6b78f04399?showInstructions=true), I was able to obtain the following graphs:
The graph on the left is the raw scatterplot of the data that was collected from the API files on April 20, 2018 at 2200 UTC, while the graph on the right is the residuals plot of that data. After observing the graphs and receiving a correlation coefficient of 0.00858, I was able to safely conclude that despite the fact that people may expect weather severity to be an important factor that heavily contributes to flight delays, this data suggests that there is no such association. In fact, this data gives convincing statistical evidence that except in extreme cases, weather has no impact on flight delays whatsoever. However, this conclusion is not entirely surprising considering that all major commercial airline pilots are all instrument rated, meaning they are not only required to be completely trained and capable to fly through areas with poor weather conditions, but in some cases they are expected to be able to do so. In fact, according to Part 121 of the Federal Aviation Regulations, all airline pilots must create an IFR (Instrument Flight Rules) flight plan before their commercial flight, ultimately preparing them for any weather condition (FAR 121).
Even though weather severity may not be the best at predicting flight delays at certain airports, it is still certainly a problem that is extremely costly to both the passengers and the airline company. According to Ann Guy, a reporter for the Berkeley News Company, flight delays are a serious issue that end up causing passengers to lose 16.7 billion dollars every year. It also costs the company itself about 8.3 billion dollars every year as well (Guy). Although my prediction in using weather to create a model that would help predict flight delays was ultimately wrong, there are plenty of other variables that can be tested with enough time. It is essential for people to continue this type of research and learn the benefits of computer programming as it is not only effective and efficient for collecting data, but it is also unbiased, which is essential to creating a representative model. I would encourage everyone to not only take a couple classes in computer science, but try to apply it in certain ways, whether it is to make an everyday task more efficient or to simulate results of a virus. Whatever it is, its purpose can play a much bigger role in modern data analysis and although my data may not have been meaningful in helping to understand what truly causes flight delays, it is certainly a step forward to help find what might prevent this problem.
“Flight Plan: VFR and IFR: Supplemental Operations.” Chapter 1.G. Federal Aviation Regulations. 2018.
Guy, Ann Brody. “Flight delays cost $32.9 billion, passengers foot half the bill.” Berkeley News. 18 October 2010, http://news.berkeley.edu/2010/10/18/flight_delays/. Accessed 18 April 2018.