BenAAndrew / sentiment-api

Python/Flask API to analyse sentiment in text

sentiment-api

https://sentiment-analyse.herokuapp.com/

A Python API using the nltk library to analyse sentiment in text

Request format

Supports both JSON and Multipart form requests.

{
    "text": "my text 123!",
    "sentiment_threshold": 0.1
}

• text: The plain text you want analysed. Non-alphabetic characters will be ignored and every word will be assessed (must be space seperated like a typical paragraph).
• sentiment_threshold (optional): Minimum sentiment value (Default is 0.1). A greater sentiment value will reduce the number of returned words only keeping those with strong positive & negative sentiment.

Returned format

Returns a JSON

{
    "words": {
        "cool": {
            "occurances": 1,
            "score": 0.3182
        }...
    }
}

Returns a dictionary of words containing each word with a sentiment greater than the sentiment_threshold. Each word dictionary contains its number of occurances in your text and it's sentiment score.

A positive value means the word has positive meaning and visa versa. A greater value means the word has greater sentiment.