Case Studies
Building High-Quality Data for Personalised AI Fitness Product Recommendations
Project Overview
A leading fitness recommendation platform in Korea is developing an AI-powered service that provides personalised fitness product and brand recommendations based on users’ goals, body metrics, and lifestyle habits. With the “Healthy Pleasure” trend accelerating and Korea’s fitness market now exceeding KRW 7 trillion (≈USD 5 billion), consumers still struggle to find products that truly match their needs. The platform aimed to close this gap by building a personalised recommendation engine covering supplements, gym wear, and fitness equipment.
Data Voucher Project Background
Despite over 4.2 million Koreans actively engaged in fitness and bodybuilding, data-driven personalisation in product recommendations remains limited. Consumers face fragmented information, while brands encounter rising marketing costs and restricted visibility. Through Korea’s Data Voucher (DV) Program, the platform collaborated with Tictag to develop a structured, high-quality dataset, the foundation for its AI-based recommendation system.
What We Did:
As part of the DV-supported process, Tictag managed the complete data lifecycle, from collection and processing to validation and AI readiness, ensuring full compliance with DV standards.
- Data Collection
- Conducted structured consumer research on fitness product usage and preferences.
- Standardise variables such as workout goals, supplement habits, preferred brands, and purchase intent into a unified metadata structure.
- Collected 1,589 responses, exceeding the initial target by 158%.
- Data Processing and Cleansing
- Removed duplicates and invalid responses, securing 1,037 valid datasets (≈65%).
- Refined data into analyzable structures for workout goals, product habits, brand preferences, and purchase intent.
- Processed and formatted datasets for direct AI model training.
- Data Validation and Quality Assurance
- Structured data for correlation analysis between fitness goals, lifestyle, and spending patterns.
- Achieved data quality levels compliant with machine learning requirements and DV verification standards.
Results & Impact
The verified dataset now powers an AI-based recommendation engine that delivers:
- Enhanced Recommendation Precision: Higher personalisation accuracy through behaviour, purchase, and brand affinity data.
- Improved User Experience: Trusted, AI-driven recommendations supported by community insights.
- Data-Driven Brand Strategy: Actionable insights for optimised marketing and consumer targeting.
✨ Power Your AI with Better Data. Accurate AI starts with verified data. Partner with Tictag now to build your next Data Voucher–grade training dataset that truly understands your users.