Frederick O'Loughlin

Sentiment Analysis

Tool to Identify key entities in text and score their sentiment

In order to prototype functionality for a Natural Language Processing text analysis app, I developed a demo application which you can access below.

The demo takes an extract of text, and uses NLP techniques built-in python in order to identify the key entities (things) in the text. It then determines a score for each of the words determined by its context, as to whether it seems to be framed positively or negatively.

As this is a demo, it is also limited, the full app I went on to develop has a more intelligent way of identifying the entities as well as the sentiment, and of classifying and summarising text. Future developments also include going beyond mere sentiment and identifying the actual emotion of the text content.

These techniques provide very powerful means of interpreting and understanding large quantities of text. Allowing effective insight into otherwise unmanageable data.

Give the demo a go here:

Let me try it!