DiscorraDiscorra

About Sentiment Analysis

An overview of sentiment scoring in Discorra — how polarity is measured, how comparative scores work, and how to interpret positive, neutral, and negative distributions across corpora.

Updated 9/12/2025

About Sentiment Analysis

Sentiment Analysis in Discorra shows you the emotional polarity of corpora — how much of the language is positive, neutral, or negative.
It helps identify not just what people are talking about, but the tone they use to express it.



What is sentiment analysis?

Sentiment analysis classifies each sentence or token into one of three categories:

  • Positive — words that express approval, optimism, or favorable tone
  • Neutral — descriptive or factual language without emotional valence
  • Negative — words that express criticism, dissatisfaction, or unfavorable tone

Definition
Sentiment analysis in Discorra measures the distribution of polarity values across a corpus and compares them across datasets.


How sentiment is measured

Discorra uses a lexicon-driven approach combined with weighting:

  1. Polarity assignment
    Tokens are mapped to positive, negative, or neutral sentiment scores.

  2. Distribution calculation
    Each corpus is broken down into percentage shares (e.g., 25% positive, 68% neutral, 7% negative).

  3. Comparative scoring
    The difference between corpora is displayed as Δ values for quick benchmarking.


How to interpret distributions

  • Positive %: The proportion of optimistic or affirming language.
  • Neutral %: The largest share in most corpora, representing factual or descriptive text.
  • Negative %: Indicates criticism, risk, or emotional tension.

⚠️ Note: High resonance does not mean similar sentiment. Two corpora may use the same vocabulary but with very different tones.


Comparative sentiment scoring

Discorra highlights differences with two measures:

  • Δ Positivity: How much more positive one corpus is compared to the other.
  • Δ Comparative: A normalized score (−1 to +1) showing relative tone.

This makes it easy to benchmark brand voice vs. audience tone, or compare competitors directly.


Why sentiment matters

Sentiment analysis helps you:

  • Benchmark brand vs. audience tone
  • Detect emerging risks when negative sentiment increases
  • Identify opportunity framing (e.g., shifting neutral terms into positive space)
  • Measure campaign effectiveness in shaping perception

Example:
In a Jazz vs. Blues comparison, Jazz shows 25% positive vs. 28% for Blues. Though close, Blues leans slightly more positive. This signals a subtle difference in tone that may influence audience perception.


Next steps


Further Reading