ExploreAssist — Personalized City Sightseeing — Empowering Your Travels with Personalized AI-driven City Recommendations!

5 min readOct 8, 2023

This project was carried out as part of the TechLabs “Digital Shaper Program” in Dortmund (summer term 2023).

Logo

In a nutshell:

Our innovative app aims to redefine travel planning through AI-driven personalized city sightseeing recommendations. By leveraging cutting-edge technologies and data from diverse sources, we create tailored travel experiences, minimizing the hassle and maximizing the joy of exploring unknown cities. Users input their preferences, and our app, backed by CHATGPT 3.5, provides intelligent, personalized suggestions for cities and sites to explore, fundamentally reshaping how we experience new places.

Introduction:

Problem: Visiting a new city can be an overwhelming experience due to numerous sights and limited time, making planning a hassle. Current solutions must provide recommendations that align well with individual preferences, leading to less-than-optimal travel experiences.
Solution: Through user input on interests such as culture, history, art, music, food, sport, and relaxation, our app generates personalized city and sightseeing recommendations, ensuring an enriching and hassle-free travel experience.
Motivation: The project appealed to our team as it engaged all disciplines — web development, data science, etc., equitably, allowing us to build a holistic solution to a common problem and making travel more exciting and fulfilling for everyone.

Methodology:

Our approach involved utilizing HTML and CSS for web development and Figma for UI & UX Design. Python powered the data science component, with data being sourced from OpenStreetMap. We employed CHATGPT 3.5, interfaced through an API, to generate intelligent travel recommendations. Our methodology allowed us to integrate diverse technologies and data sources, building a cohesive and powerful travel recommendation engine.

Furthermore, we implemented the SCRUM framework in our process to incrementally move forward with our app. One of our members suggested Azure DevOps but ultimately decided to utilize GitHub for our tasks and have weekly scrum calls with everyone and a call with the data scientists every two weeks. This was especially needed since we mainly worked remotely and tended to slow down. Working in a team of six, regular push into the GitHub repository may result in many conflicts. That’s why we created branches for each discipline to prevent multiple conflicts.

A server was also set up on which everything should run. While the implementation was successful, we had difficulties transferring the users’ data to the AI and the result to the website, as seen in the branches “websiteToAi” and “aiToWebsite”.

Data Science: In Data Science, the cornerstone of our application, the primary challenge was to synergize diverse data sources effectively. We relied on Python, a versatile language known for its efficacy in handling data-related tasks, to manage and process data from OpenStreetMap and other datasets available on platforms like Kaggle. While we successfully retreated the data from OpenStreetMap, we implemented an advanced AI model for generating personalized travel recommendations, since the data from OpenStreetMap wasn’t enough for our goal. We incorporated CHATGPT 3.5 to provide intelligent, contextual suggestions based on user preferences. Integrating this sophisticated model required meticulous fine-tuning and optimization to ensure seamless interoperability with our system. We overcame complexities associated with interfacing different components and established a robust API connection between our application and CHATGPT 3.5, allowing real-time, accurate travel recommendations.
Web Development: The Web Development aspect focused on creating a user-friendly and responsive interface. Given the diverse range of user devices, designing a universally compatible and seamless interface presented its challenges. Utilizing HTML and CSS, we built a lightweight yet robust structure, ensuring optimal performance and user experience across devices. This process was vital to optimizing the web components for various screen sizes and resolving compatibility issues to ensure uniform user experience. By implementing responsive design principles and rigorous cross-browser testing, we harmonized design aesthetics and functionality, allowing users to interact with our app effortlessly.
UX/UI Design: For UX/UI Design, the challenge lies in translating user needs into an intuitive and engaging user interface. Using Figma, we crafted design prototypes focusing on user-centric design principles. We prioritized simplicity and clarity, ensuring users can easily navigate the app and access its features. User feedback was crucial in refining our designs, helping us identify and rectify usability issues. We enhanced interface intuitiveness and user satisfaction by fostering an iterative design process and incorporating user insights. Our meticulous attention to user experience details resulted in a design that is not only visually appealing but also elevates the overall user interaction quality with the app.

Results:

Our application, culminating in multifaceted technological integration and innovative design, is a pioneering solution in personalized travel recommendations. It integrates sophisticated AI to analyze user preferences and deliver highly personalized travel suggestions, spanning many cities worldwide.

The central outcome of our project is establishing a user-friendly platform offering intelligent, custom-tailored travel recommendations. The seamless synergy between data science, web development, and UX design has enabled the realization of a tool that redefines travel planning, rendering it a hassle-free and enjoyable experience.

Our application promises a transformative impact on the way individuals plan their travels. It eliminates the overwhelming aspects of travel planning, offering a streamlined, intelligent approach to exploring new cities. By fostering a more enriched and hassle-free travel experience, our app has the potential to redefine norms in travel planning and exploration.

First, we need to finish the changes in “websiteToAi” and “aiToWebsite” with some debugging. Second, we must adjust the website’s frontend to the set UX/UI Design requirements. After all that, we are at the threshold of our journey and eagerly anticipate user feedback to refine our application. Immediate developments include enhancing user interaction, refining recommendation algorithms, and incorporating additional features based on user insights and needs. We aspire to evolve continually, optimizing our app to meet the dynamic demands of modern travelers.

We aim to expand our app’s capabilities to offer comprehensive travel planning solutions. We also aim to enable users to receive personalized recommendations and create, modify, save, and share intricate travel plans. By integrating collaborative features, we envision our app becoming a central hub for travel enthusiasts, where users can collectively contribute to creating unparalleled travel experiences, keeping our app at the forefront of travel innovation.

In conclusion, our application marks a significant stride in personalized travel recommendation with its intelligent, user-centric design and advanced technological integration. We are optimistic about its potential to shape future travel experiences, and we are committed to its continual refinement and expansion to meet the ever-evolving needs of travelers worldwide.

GitHub repository (or similar):

https://github.com/TechLabs-Dortmund/Personalized-City-Sightseeing/tree/main

Figma Design: https://www.figma.com/file/odXdJBYPwVIg2ORhLOQaL6/App-City-Recommendation?type=d esign&node-id=273%3A560&mode=design&t=WGAR4fbs0caDNnfI-1

Team members:

Data Science: