Inspiration
Marketers often scramble to catch trends after they’ve peaked, so I wanted to build a tool that lets anyone spot hot topics the moment they emerge and instantly plan content around them.
What it does
SparkTrend is a simple, two-page web app that helps turn live Google Trends data into actionable marketing insights.On the landing page, you learn what’s trending and how SparkTrend works. On the chat page, you just type a topic and SparkTrend pulls the top 10 most-searched terms from Google’s public Trends dataset. This allows businesses to create marketing campaigns tailored to the world's most relevant concepts.
How we built it
I created a multi-agent pipeline with three key steps: Parameter extractor to parse a user’s request into topic, region, and time window. SQL generator/executor that queries Google’s public BigQuery Trends dataset for the latest search terms. Advice agent that ranks the results and produces concise marketing recommendations. And then implemented the front-end using React and back-end using FlashAPI for the project.
Challenges we ran into
I had a very hard time running the backend properly because it was facing problems in producing outputs, which led me to be really stressed.
Accomplishments that we're proud of
- Built a working multi-agent system from scratch that dynamically queries real Google Trends data.
- Designed a clean, user-friendly chat interface where marketers can interact with live search trends in plain language.
- Learned how to orchestrate agents so that each step passes state smoothly to the next.
What we learned
I learned how to build a model with 3 agents, and implement SQL queries
What's next for SparkTrend
I would hope to sharp up the recommendation for content, marketing ideas for the most searched terms, so the platform can take a step up in helping everyone produce more engaging content.
Log in or sign up for Devpost to join the conversation.