-
-
The homepage of our website
-
Jhon F Kenedy: KJFK Safety analysis
-
AI Questionaire
-
Question in the chatbot: "What is the average windspeed in knots?"
-
Question in the chatbot: "Can hailsorms affect aircraft preformance and is it dangerous to fly in them?"
-
Toronto Pearson: CYYZ Safety analysis
-
How to use
Domain.com
We registered the domain pleasedontcrash.tech with domain.com.
Inspiration
As one of our team members was flying in a plane with his mentor, they heard a story on the radio: a few hours ago a team of pilots got iced (developed ice on the wings of their plane) right before they landed. Although they landed safely, one pilot got injured and it was a scary and dangerous experience. We want to prevent such events from happening, by making a site that warns pilots about the weather conditions in different regions across airports using METAR.
What it does
FlyCon is a website that identifies the safety of flying on a route on a particular date and time using the Meteorological Terminal Air report (METAR) data from METAR-TAF.com API. This website rates weather conditions in 5 categories, Perfect, Safe, Risky, Dangerous, and Extremely Dangerous. Furthermore, our website includes an AI in which you can ask any questions regarding the weather dangers and it would give you an elaborate explanation of what exactly poses any risk if there is some sort of danger. The AI can also tell you the ICAO code for every airport. For example, you could search in the AI search bar for the ICAO code for the Pearson Airport, and it will give you the corresponding ICAO.
How we built it
We built the backend with Python and Flask. The front end we built with HTML and CSS using Bootstrap templates. In Python we implemented 3 API’s: the Co:here API for our AI chatbot, we used the METAR-TAF API for real time weather information from airports worldwide and we also used the NAV Canada Area Charts for graphical weather maps published by the government of Canada.
Challenges we ran into
We ran into numerous challenges over the course of this hackathon. One major challenge was finding which METAR parameters were safe and which weren’t. For example, a wind speed of 11 knots is risky if there are no other parameters, but is only rated slightly more dangerous than the safe category. However, a speed of 10 knots without other parameters would be in the safe category. It is a slight difference but the categories they are ranked in are different. Now with the other parameters the ranking becomes a lot harder and even with safe wind speed, could be ranked in the risky or even the dangerous category. Another challenge we ran into was allowing the website to be viewed on devices with smaller widths, such as smartphones. We implemented a ‘hamburger’ menu to allow the website to be viewed on small devices, without affecting the functionality of the website. Prior to this feature being added, the website would be virtually unusable on small screens. Another big problem was implementing AI. Since it was our first time doing this, we were not entirely sure how we should add it. The last problem we ran into was hiding the key to the METAR API. When we published our code to Github, we would not want there to be any sensitive information. We hid the API key by using GCP’s Secret tool. Other minor challenges included agreeing on some of the text added to the website.
Accomplishments that we're proud of
In this project we learned Co:here and implemented it for the first time. One of the accomplishments we are very proud of is connecting the METAR API to the website so it's accessible to you and us. Also, we are proud that we were able to manage our time well and work well as a team to successfully complete this project.
What we learned
During the making of this website, we learned how to use Flask and Bootstrap to make websites, and how to create pipelines between APIs and websites. As a result, we were able to complete the project in a shorter period of time and add more functions to it.
What's next for FlyCon plans to prevent weather related aircraft crashes.
The next step for the website is to make it look more professional and spend some more time on the layout and design of the site itself. Not only that, but we would also implement a system to make different units of measurement into the same (to avoid confusion), as an identical-looking METAR may mean different things depending on where you’re flying. But most importantly we want to sell our website to an insurance company. An insurance company would be interested in the chances of something like a plane crash happening and therefore, will use our website to get a more accurate prediction. Finally, we would add keyboard shortcuts to allow users to go from one page to another easier.
Log in or sign up for Devpost to join the conversation.