Analytics
Sentiment Score
Scores text from −1 (negative) to 1 (positive) with a built-in English word list, and labels it Positive, Neutral or Negative.
When to use it
Use it to sort reviews, survey comments or tickets by tone, or to track sentiment over time.
In a pipeline
File InputSentiment ScoreSummarize (Multi-Aggregate)
Settings
| Setting | Type | Required | Default or hint |
|---|---|---|---|
| Text column | select | Yes | Select column… |
| Word list | select | — | en |
| Score column name | text | — | Default: Sentiment |
| Add a Positive / Neutral / Negative label | checkbox | — | true |
Example
Review sentiment
- Text column=Review
| Review |
|---|
| Great service, really fast delivery! |
| The parcel arrived on Tuesday. |
| Terrible, the box was broken and nobody helped. |
| Review | Sentiment | Sentiment Label |
|---|---|---|
| Great service, really fast delivery! | 0.796 | Positive |
| The parcel arrived on Tuesday. | 0 | Neutral |
| Terrible, the box was broken and nobody helped. | -0.791 | Negative |
Worth knowing
- Word-list based, not a language model: “not”, “never” and friends flip the next words, “very” strengthens them and words after “but” count more — sarcasm and context are beyond it.
- Scores from −0.05 to 0.05 are Neutral. Blank text gets a blank score.
- English only for now.