City Pulse

Inspiration

As a travel enthusiast, I understand the challenge of crafting a personalized trip; it's time-consuming and complex. City Pulse simplifies this by connecting travelers with the city's essence, from popular sights to hidden gems, tailored to individual preferences in food, art, history, or entertainment. Each journey is unique, ensuring a memorable experience.

What it does

City Pulse requires three inputs: city name, user's needs, and preferred language. It generates a collage of symbolic city images, recommends the top three must-see attractions, offers a brief city introduction, and includes a chatbot to tailor itineraries to user preferences and timelines.

How we built it

I researched the functionality of pary rock through medium post, youtube and aws documents. The party rock is pretty straightforward and no code.

Challenges we ran into

Sometimes, for foreign cities, the LLM experiences issues with hallucination; Images we generated are not consistent with prompts.

Accomplishments that we're proud of

I am impressed by how much I can achieved in such a short time. And I will use this app for my next trip. And i will share this app to my readers on medium.

What we learned

I learned the iterative improvement of prompt engineering for context generation; and use prompts to guide QA; orchestrate different Gen AI application using @variables

What's next for City Pulse

Reduce hallucination through few-shot learning and improve image generation. Link to APIs to help users find the best deals on flights, cars, hotels, weather, and local food, restaurants, among others. The possibilities are endless.

Built With

• amazon-web-services
• apigateway
• chatgpt
• dynamodb
• lambdafunction
• python
• sklearn
• tensorflow