trends | Devpost

Inspiration

The overwhelming amount of content on social platforms inspired us to create a tool that simplifies the discovery and organization of trends. We wanted a platform where users can follow topics of interest, engage in meaningful discussions, and archive impactful social moments for posterity—all while empowering users to uncover nascent trends before they go mainstream.

What it does

Trends is a curated platform for tracking, summarizing, and engaging with trending social topics. It aggregates posts from platforms like X, categorizes and summarizes them using GPT-4o, and organizes them into thematic boards. Users can interact with these boards, upvote, comment and explore how discussions evolve over time. The source data of tweets are permanently archived using Arweave, creating a historical record of social engagement.

How we built it

• Frontend: Built using Next.js with Tailwind CSS for a sleek, responsive, and user-friendly interface. Have to use client side rendering to deploy to arweave. Exploring ways to enhance speed of data fetching and rendering. • Backend: All data is stored on AO. Using Arweave to store the tweets permanently, so the backend summarises can be recreated with more powerful models. • AI Models: Utilized GPT-4o and GPT-4o-mini to categorize and summarize tweets dynamically.

Challenges we ran into

Data Aggregation: Parsing and cleaning data from X in real time while ensuring meaningful content was a challenge. Apify really helped with this.
AI Categorization: Balancing accuracy and speed when categorizing and summarizing tweets using GPT-4o and GPT-4o-mini. Had to experiment with the prompts. Still a lot of scope for improvement.
User Experience: Designing a UI that simplifies complex trends while making it visually appealing and engaging. Dealing with a large amount of data on AO. More scope to create better data structures and data handling as the amount of data grows.
Deploying to the Permaweb: Unable to use cutting edge Next.js features like server components and dynamic metadata. Had to revert to client side rendering and yet faced problems in deployment. ArLink was helpful in deploying.

Accomplishments that we're proud of

• Successfully built a system that uses AI to dynamically curate and summarize social trends. • Integrated Arweave to provide permanent storage for significant trends and discussions. • Developed a new type of social platform that can leverage existing distribution (X etc.). Had a lot of meaningful discussions with members of the Arweave community discussing future plans of tokenized engagement model to gamify interactions and make discussions more meaningful, AI agents, expanding range of topics covered etc. • Created an intuitive and visually compelling mobile responsive UI/UX in a short timeframe.

What we learned

• The power of AI in transforming unstructured social data into curated insights. • The potential of decentralized storage (Arweave) to preserve critical social engagement and discussions. • The importance of designing with the user in mind to simplify interactions with large volumes of data. • Several ideas around gamification (Trend Tokens) and AI that can enhance user engagement in community-driven platforms.

What's next for trends

• Community Features: Add collaborative tools like co-curated boards, polls, and shared archives. • User Profiles & Analytics: Provide users with a profile and insights into their engagement and trends they’ve participated in. • Gamification using Trend Tokens: To enhance user engagement in community-driven platforms. • Expanding Sources: Curating more topics to attract users from across web3. • Feedback: Talking with more community members to create a future roadmap.