TrendPulse is an AI-driven trend intelligence platform designed to help travel creators make better content decisions—not based on intuition, but on real-time data.

The project focused on understanding how large-scale trend data is collected, processed, and delivered to users in a scalable cloud environment, while exploring how UI can surface complex backend insights in a clear, actionable way.

This project was less about polished UI and more about learning how product design, data pipelines, and cloud architecture work together to support real-world services.

Travel creators often rely on instinct or outdated tools when deciding:

• what destinations to cover
• which topics or keywords to use
• when to publish content

This leads to four major problems:

• Guesswork decisions instead of data-backed choices
• Missed timing when trends peak before content is published
• High production risk due to costly travel content
• Keyword blind spots that limit reach and discoverability

TrendPulse provides creators with:

• real-time trend signals
• sentiment and mood analysis
• AI-assisted content ideation

so they can answer three key questions:

• Collects trend data every 2 hours
• Highlights rising destinations and topics before they peak
• Reduces guesswork in destination and topic selection

• Analyzes how people emotionally respond to destinations and trends
• Helps creators align content tone with audience sentiment
• Supports more emotionally resonant storytelling

Converts processed trend data into:

• short-form content ideas
• hook phrases
• thumbnail concepts
• BGM suggestions

Bridges raw data and creative execution

• Allows users to ask questions directly about trends
• Responses are powered by the most recently processed data
• Demonstrates how backend intelligence can be surfaced conversationally

TrendPulse differentiates itself by combining:

• pre-creation trend insights
• keyword & hashtag guidance
• content planning support
• sentiment & emotion analysis

Unlike existing tools, TrendPulse focuses on timing + emotion, not just keyword optimization.

The system is divided into two independent layers:

1. Data Pipeline Architecture

• Collects data from sources like YouTube and Reddit
• Processes trends on a scheduled basis (every 2 hours)
• Handles sentiment analysis and data enrichment
• Designed so failures never affect user experience

2. User-Facing System

• Serves fast, responsive user interactions
• Queries only processed and stable data
• Allows independent scaling from the data pipeline

This architecture ensured:

• fast UI response times
• reliable user experience
• scalable growth without reworking the system

Through this project, I learned how:

• cloud-based architectures drastically reduce upfront costs
• auto-scaling supports growth without operational overhead
• MVP systems can be designed to stay cost-efficient while scaling

TrendPulse demonstrated that a data-heavy AI service can be:

• operationally lightweight
• financially scalable
• realistic for early-stage products

• Participated in early UI design and feature definition
• Collaborated closely with backend and data engineers
• Learned how AWS services (Lambda, S3, OpenSearch, Bedrock) connect in real products
• Focused on understanding backend constraints and designing UI that respects them

• Good UI depends on how data is processed behind the scenes
• Architectural decisions directly shape user experience
• Product designers benefit from understanding backend systems—not just interfaces

TrendPulse strengthened my ability to design system-aware interfaces, not just screens.