Sophia NLU Engine - Usage Statistics and Reporting

Gain deep insights into your NLU implementation with Sophia's comprehensive analytics system, tracking key linguistic patterns and performance metrics in real time.

Logging & Storage

Sophia’s robust logging system ensures all usage data is captured and stored efficiently:

  • Multiple configurable log filters and targets
  • Support for text files, Redis, SQLite, MySQL, and PostgreSQL
  • Automatic log rotation and management
  • User-level tracking with UUID support

Analytics Features

Analyze and visualize your NLU data with flexible, powerful tools:

  • Real-time performance monitoring
  • View data by hour, day, week, month, quarter, or year
  • Filter by timescale (all results matching a filter) or by specific elements (e.g., word, POS tag) within a time period
  • CLI view resembling a MySQL prompt, or export to HTML, CSV, or JSON

Key Metrics

Sophia tracks a wide range of metrics to provide actionable insights:

  • Frequencies of POS tags and tag categories
  • Word and multi-word expression (MWE) usage
  • Verb + noun pairs within the same phrase
  • Input volume and token usage
  • Named entity recognition stats
  • Category usage patterns
  • User engagement metrics
  • Response time analysis
  • Selector performance tracking
  • Model confidence scoring
  • Traffic patterns (hourly, daily, weekly, monthly, quarterly, yearly)
  • Integration performance metrics
  • Error rate monitoring
  • Custom metric definition