Pulse

A streamlined tool which scrapes news websites to detect natural disaster headlines and sends targeted emails to subscribed companies, notifying them about events in relevant geographical areas.

Comment 1

Inspiration

Natural disasters impact millions of people globally each year. In 2023, for instance, approximately 93.1 million individuals were affected by such events . Over the past few decades, the average annual death toll from natural disasters has ranged between 40,000 and 50,000. One common theme that often leads to more casualties is response times being slow. Therefore, Pulse wants to emphasize on the importance of a quicker response time, as it will allow casualties to be reduced drastically.

What it does

Automated Scraping: Using a weather API to detect locations where extreme weather events are happening. Regularly scrapes news articles for extreme weather alerts using a Python backend. Targeted Alerts: Sends customized emails to those registered on the mailing list based on the type of natural disaster and recipient’s location. User Subscription: Simple front-end interface for companies to join the mailing list, providing their email, company type, and location.

How we built it

We looked at different themes and identified the ones we were most interested in (cybersecurity and extreme weather) We each identified personal experiences that could help us identify potential pain points and problems to tackle (ending up deciding on supply chain problems) Identified technologies we were all proficient in, and designing the architecture of the product around that technology (mainly Python) Assigned tasks to each person and kept up good communication throughout to ensure we were working towards the same direction Finalized frontend and backend and integrated them all together during the last few hours of the hackathon

Challenges we ran into

One challenge we encountered was in scraping the articles; going with BeautifulSoup and trying to match keywords to article titles was largely inconsistent (e.g. we saw an article with The Hurricanes as a sports team name), while asking ChatGPT to look at the HTML code for the website was highly risky as it consistently hallucinated. In the end, we decided on a combination of both, matching keywords to article titles, then verifying whether they were actually extreme weather conditions using a ChatGPT call. Another challenge we encountered was deploying the fullstack application in such a short time. Attempting to apply our new-found Next.js skills in addition to learning how to connect that to a backend that sent requests to MongoDB (also a new database) was a huge difficulty curve. The team also faced challenges related to a consistency of knowledge throughout all members, in some stages, not all the members were aware of the next step leading to time loss.

Accomplishments that we're proud of

We are proud of being able to develop a working demo with our limited resources and time, as well as performing at a high level as a team. As it is our first Hackathon, we made significant progress as hackers and also cemented our interest into the tech world.

What we learned

We developed technicals skills relating to both front-end and back-end web development. We also improved our use of Python modules such as BeautifulSoup and Requests. We also used ChatGPT API to automate a mailing process which adapts to the content being fed.

What's next for Pulse

The step for Pulse is to implement a way to assess the severity of an extreme weather event. This would allow companies and non-profit organizations to better manage their resources and control their supplies. Another future add-on that can be implemented is to include a heat map on the web app. This heat map would show the natural disasters happening in a certain region and allow for quicker response.

Share this project:

Updates