The vocabulary of the Sentiment Analysis API

The 4 fields and concepts you'll meet in the response — defined in plain English, each with a real example value.

4 terms
Core Concepts2

Sentiment Score

A numeric value representing the emotional tone of text, typically ranging from negative to positive.

Sentiment scores quantify subjective opinions expressed in text. Most scales range from -5 to +5 or -1 to +1, where negative values indicate unfavorable sentiment and positive values indicate favorable sentiment. Zero represents neutral text.

ExampleThe text "I love this product!" might receive a sentiment score of +4, while "This is disappointing" might score -2.

Polarity

The direction of sentiment—positive, negative, or neutral—without regard to intensity.

Polarity classifies text into broad categories. While sentiment scores provide granular measurement, polarity provides categorical labels useful for filtering, routing, and simple analytics.

ExampleBoth "This is good" (+1) and "This is absolutely amazing!" (+5) have positive polarity, despite different intensities.

Metrics1

Comparative Score

Sentiment score normalized by text length, allowing fair comparison between texts of different sizes.

The comparative score divides the total sentiment by the number of words analyzed. This normalization enables comparing a short tweet to a long review. A comparative score of 0.3 means the text averages 0.3 sentiment points per word.

ExampleA 5-word tweet with sentiment +2 has comparative score 0.4. A 100-word review with sentiment +20 also has comparative score 0.2, indicating the tweet is more intensely positive per word.

See these fields live. Run the Sentiment Analysis API free — no card, no signup wall.

Scaling up?

Volume pricing, custom SLAs, and dedicated support for high-traffic teams.

Contact sales