Skip to main content
✨ NEW
Dashboard Analytics
To view statistics, patterns, anomalies, etc., at a glance for the audios associated with a Configuration.

Analytics Dashboard

The Analytics Dashboard provides powerful, AI-driven insights from your transcription data. Analyze trends, discover patterns, identify anomalies, and gain a deep understanding of your audio data at scale.
Analytics are generated for transcriptions processed under a specific configuration.

Analysis Sections

The dashboard is organized into thematic sections, each providing different insights:

Synthesis

AI-generated executive summaries

General Distribution

Core metrics and distributions

Quality & Confidence

Data completeness and anomalies

Extraction Fields

Field-by-field analysis

Temporal Analysis

Trends over time

Relationships

Correlations and co-occurrences

Synthesis

AI-powered summaries to understand your data at a glance.
  • Executive Summary: Highlight key findings, temporal trends, and opportunities or risks for decision-making.
  • Qualitative Summary: Identify recurring themes, issues, and keywords in the audios based on the qualitative analysis of textual fields.

General Distribution

Fundamental metrics about your audio files.
  • Duration
  • Speakers
  • Languages

Audio Duration Distribution

Displays:
  • Min: Shortest audio file length
  • Average: Mean duration across all files
  • Max: Longest audio file length
  • Bar chart: Distribution histogram showing frequency of different duration ranges
Insights:
  • Identify typical conversation length
  • Spot unusually long/short calls
  • Understand variance in your data

Quality & Confidence

Assess the completeness and reliability of your extracted data.
  • Completeness
  • Audios to Review
  • Anomalies

Data Completeness

For each extraction_field, a progress bar shows the percentage of transcriptions where a valid (non-null) value was successfully extracted.Use this to:
  • Identify which fields are reliably extracted
  • Find fields that need description improvements
  • Understand data gaps in your conversations

Extraction Fields Analysis

Detailed, field-by-field breakdown of your extracted data. Each extraction field gets its own card with visualizations tailored to its data type.
  • String & Array Fields
  • Number Fields
  • Boolean Fields

Text and List Fields

For string and array type fields (e.g., “Category”, “Mentioned Products”).Value Frequency ChartA bar chart showing the most common values and their counts.Insights:
  • Identify most common categories/topics
  • Prioritize product features to develop
  • Understand customer pain points
  • Track trends in conversation topics

Temporal Analysis

Understand how your data evolves over time.
  • Trend Chart
  • Activity Heatmap
A multi-axis chart plotting the evolution of metrics over time.Displays:
  • Line charts: Numerical field averages over time
  • Bar chart: Volume of transcriptions per time period
  • Multiple Y-axes: Different scales for different numerical extraction fields metrics
Customization:
  • Select which fields to display
  • Choose time granularity (daily, weekly, monthly)
  • Filter date range
Insights:
  • Track improvement over time
  • Correlate volume with quality
  • Identify seasonal patterns
  • Measure impact of changes (new features, policy updates)
Use cases:
  • Monitor customer satisfaction trends
  • Track support quality improvements
  • Seasonal demand analysis
  • A/B testing results validation

Relationships & Patterns

Discover hidden connections between data points.
  • Correlations
  • Co-occurrence

Numerical Field Correlations

A heatmap matrix visualizing the correlation between all numerical extraction fields.Heatmap colors:
  • Strong positive (blue): Values move together (0.7 to 1.0)
  • Weak/no correlation (white): No relationship (-0.3 to 0.3)
  • Strong negative (red): Values move oppositely (-1.0 to -0.7)
Interpretation:
  • Price ↔ Duration: 0.68 = Strong positive correlation
    • Longer conversations → Higher prices
  • Rating ↔ Duration: -0.05 = No correlation
    • Conversation length doesn’t affect rating
Use cases:
  • Identify factors that impact key metrics
  • Understand causation vs. correlation
  • Validate hypotheses

Need Help?