Enhancing Data Analysis with AI
Gained experience in integrating image and text data for comprehensive SNS analysis, utilizing various data sources through advanced AI-driven techniques.
Developing Brand Communication Strategies
Strengthened data-driven decision-making skills by proposing effective brand communication strategies based on thorough data analysis.
Teamwork and Collaboration Skills
Built experience in collaborating with team members from diverse academic backgrounds, leveraging individual expertise and complementing each other to achieve project goals.
In corporate marketing strategies, analyzing consumer behavior and preferences is crucial for establishing effective brand communication strategies.
There have been no systematic examples of consumer trends being analyzed through SNS content analysis.
It is necessary to integrate AI technology to analyze the direction of relationships between consumers, products, and brands, and to provide creative solutions for brand communication based on these insights.
Analyze how consumers perceive and use brands and products by leveraging SNS data.
Define the personas of current and potential consumers through the analysis of SNS users, and develop solutions that enhance the strategy of brand communication and advertising by understanding the relationship between consumers, products, and brands.
The goal is to optimize brand communication by identifying how consumers discuss specific products on SNS, and by understanding consumer personas and brand perception.
Target SNS: Instagram
Target Products (Keywords): Galaxy Watch 4, Apple Watch 6, Z Flip 3
Data Collection Method: Collected image and text data through web crawling
Data Preprocessing: Removed duplicate and irrelevant posts (ads, religious content, etc.)
Data Scale: Galaxy Watch 4 - 1,017 posts / Apple Watch 6 - 1,144 posts / Z Flip 3 - 965 posts
Used the YOLO object detection model to identify objects within images and analyze the contexts in which consumers are using the products.
Result
Galaxy Watch: In the image on the left, the detection of a cup, person, and dining table indicates that the product is being used in a café setting.
Z Flip 3: In the center image, the detection of a person, kite, and backpack suggests that the product is being used outdoors.
Apple Watch 6: In the image on the right, the detection of a keyboard, cup, and cell phone suggests that the product is perceived as an everyday item.
Examples of Image Analsis
Used the Kakao Hangul Analyzer III (khaiii), a Korean morphological analyzer, to analyze Instagram post texts.
Grouped words into similar categories to analyze how consumers use words related to the products.
Result
Everyday Life: cup (56), chair (30), dining table (18), cake (19), coffee (27), daily life (12)
Office: book (31), chair (30), tie (33), mouse (18), laptop (16), keyboard (12)
Strap: strap (431), leather (133), metal (94), strap change (58), silicone (54), band (42)
Case: case (234), card (139), hard (138), bumper (137), design (121)
Identified users' actual usage environments and key factors through Instagram.
Users use the products comprehensively in various environments such as everyday life and office settings.
We found a high interest in additional accessories like straps and cases.
Limitation & Future Work
Data collection was limited to specific products and platforms, making generalization difficult.
Additionally, data collection through crawling may face restrictions depending on policy changes.
After improving each model's performance, they can be integrated into a single platform for real market applications.
Various models beyond YOLO and khaiii can be explored.
Real-time model results can be utilized to set brand targets and provide solutions for tailored advertising strategies.