Inspiration
As the world shakes off the last vestiges of its travel slumber, the thrill to once again roam free and explore every corner is undeniably electric. Here at Patatas Bravas, we share that excitement, and we're thrilled to offer you the travel tool you've always dreamed of. We've taken a bold, unprecedented approach to travel—Patatas Bravas is a versatile, modular AI-powered tool that's set to revolutionize the industry. Gone are the days of juggling multiple sites for your transportation, accommodations, or event bookings while trying to connect with like-minded explorers. Patatas Bravas is more than a tool; it's an innovative platform designed to weave together the threads of adventure and camaraderie into a tapestry of unforgettable journeys. So, let's get right to the heart of it: "Unlock the Universe: Navigate the Globe with Fellow Visionaries in Pursuit of Wonder!" With Patatas Bravas, your next great adventure is just around the corner. Join us, and let's make travel magical again.
What it does
When it comes to attending spectacular events, planning a trip can often feel like navigating a labyrinth. From hopping between multiple websites to snag the best travel connections to consulting various platforms for the ultimate experience, the journey planning process can be daunting. Seizing this challenge, we at Patatas Bravas felt a powerful urge to make a mark on the industry. Our vision is to redefine how travel time is perceived, introducing a new era—much like the transformation brought by innovations like GPT. Our mission is straightforward: to make travel effortlessly accessible to everyone. We're especially focused on those who have hesitated to take the final steps needed to maximize their travel experiences. With Patatas Bravas, we're not just simplifying the process; we're transforming it, ensuring that more people than ever can explore the world with ease and enthusiasm and make connections for life.
How we built it
With a lot of magic and love! :) We had the user experience in the forefront about curating a travel plan with cities that match their interests along with the company of like-minded people. We used publicly available datasets with routes of flights and trains to model the connected cities in the world, which we use to retrieve the possible routes that a user can take from a given Source and Destination. We used the vector search and vector database platform IRIS Vector Search to calculate embeddings of major cities, topics of interests for an user and a user bio embedding for each user as well. We then again use the platform to rank the cities for each user based on similarity between the city and interest embeddings as well as each user to other users based on the similarity between the bio embeddings. For a user search on Source city and Destination city, we return the sorted routes based on the similarity between the intermediate layover cities and the user interests and also show Events happening in the cities for the users to attend. If there are some travel routes that other users on the platform are also taking, and these users share some common interests calculated via the embeddings of the user bios, we provide an option for the user to connect with them if they want. We are also concerned with Sustainable travel, and so we also give suggestions for train routes for less CO2 emissions but we also are transparent to show the extra time that will be needed if trains are taken instead of flights. Finally we provide options for accommodation ranging from low budget -> budget friendly -> luxury for the complete travel experience with Patatas Bravas.
We built a very intuitive ReactJS frontend alongside a Flask (python) backend. For the current stage, We managed entire functionality with http-requests. We provided a map with route visualizations, and a nice origin/destination selection feature. We provided an autocompletion function as well, yet we decided to hide it from the UI for now due to API limits we've reached. Overall, we believe we managed to present a good amount of our work in our full-stack app!
Challenges we ran into
From the get-go, we understood that if creating Patatas Bravas were easy, it would have already been done. One of our earliest hurdles was gaining access to open-source data and securing the necessary API keys—a challenge that proved both daunting and enlightening. Deploying a tool that lives up to our high standards requires an immense amount of work. Our team, deeply rooted in theoretical mathematics and data science, initially found ourselves at a crossroads when it came to aesthetic web design. Creating an engaging and visually pleasing frontend presented a steep learning curve that we had to quickly ascend. Despite these challenges, our commitment to transforming the travel experience drove us to expand our skills and push the boundaries of what we could achieve. Each obstacle has been a stepping stone to greater innovation and deeper understanding.
With InterSystems integration there were a few issues setting up the connection between the dataframe and the SQL Alchemy interface of Iris, which we eventually resolved. However, one of the ideas we had with using Llama index doesn't work due to an issue with the global handler of llama_core.
Accomplishments that we're proud of
We're proud of so many things we did over this weekend. Firstly, planning our trip to Spain and managing everything! It was our first time here and we couldn't be more grateful:) For the hackathon itself we're proud of how we came up with such a unique idea to create a memorable travel experience. Something I feel everyone will actually use and enjoy, cause we sure think we do. The User Interface flow is actually smooth and has all the complex features inbuilt into it that we're so proud of. Since we're going for the InterSystems challenge too, navigating the complexity of Iris vector database and perfectly fitting it to our use case (with interest based similarity of users as well suggesting different routes based on similarity of cities) is something we're so so proud of. Huge shout-out to the InterSys mentors for engaging conversations throughout the weekend. Finally the images we generated and the catchy names we've used throughout is something we think is to be proud of as well!
What we learned
Time passes fast. Like REALLY fast. Luckily we had 36 hours to code to our hearts content and we came up with a unique full stack project. Since we had a lot of interesting challenges and our team was divided on which one to take, we learnt how to balance creativity and implementation and through a long discussion come to a mutual conclusion on something we'd all love to be a part of :)
Since we're all travellers we also seek good user experience so we learnt how to stitch together a great user interface that people would actually want to use.
All this and so much more, and what better place to learn and experience this than HackUPC :)
What's next for potatoes-patatas
Patatas Bravas has made significant strides since our initial 36-hour development sprint. Looking forward, we plan to integrate more real-time data feeds and comprehensive travel schedules. Our aim is to ensure that Patatas Bravas becomes the most complete and user-friendly tool in the market.
Additionally, we're focused on refining the AI that powers our platform, enhancing its ability to automatically plan personalized sightseeing trips. This means our users can enjoy perfectly tailored travel experiences without the hassle of planning every detail themselves.
As we continue to expand and refine our offerings, our commitment to the innovative spirit that birthed Patatas Bravas remains unwavering. Our goal is clear: to revolutionize the travel planning process, making it as seamless and enjoyable as the journey itself while doing something good for the environment.
Built With
- css
- flask
- html
- huggingface
- intersystems
- iris
- javascript
- leaflet.js
- llama-index
- python
- react
- transformers
- vector-search
Log in or sign up for Devpost to join the conversation.