We build a unified database from millions of media records and analyze the essentials at maximum speed
The unified database designed on top of Elasticsearch, one of the world's leading search engines, enables rapid search and analysis even within massive datasets containing millions of records.
Unified Media Data Lake
All unstructured data, including news and social data, is stored in an integrated data lake. Elasticsearch real-time indexing enables search and analysis without delay even at large scale, while providing the flexibility to grow as data volume increases.
Time-Series Keyword Analysis
Time-series analysis algorithms detect patterns in changes to keyword mention volume.
AI also supports spike detection to automatically identify surges in public attention, helping teams avoid missing the critical window when an issue emerges.
NLP Related-Term Analysis
Using NLP-based co-occurrence analysis, the system visualizes semantic connection networks between keywords.
This supports topic modeling that reveals hidden issues and core agendas, making it easier to understand the specific context behind public sentiment.
Deep Learning Sentiment Classification
A deep learning model based on a modern Transformer architecture performs contextual embedding to interpret the meaning of words within context.
This enables high-accuracy sentiment classification that can even understand nuance and sarcastic expressions.