Vox - Making voiceprints out of text content.
0 likes359 views
Tagging text with parts of speech tokens enhances content query flexibility and provides deeper insights through grammar and graph analysis. The document details an analysis of Nokia's social media conversations over three months, visualizing the brand's 'voiceprint' by measuring tone and frequency across seven parameters. A report is available at a specified URL, and feedback is welcomed.
Vox - Making voiceprints out of text content.
- 1. One of the many advantages of tagging text with Parts of Speech (PoS) tokens is that it makes the content much more pliable for querying. The assumption, and the finding, was that when it comes to text analysis, grammar knows best. And that grammar and graph work very well together to give you more insights into the data. In this instance, we analysed Nokia’s three months of Twitter|FB conversation data, and graphed them across seven brand parameters, after ascribing “tone” to the content. (Instead of sentiment, we go by tone.) By plotting Frequency (high-low scale), with Tone (positive-negative scale), we publish a “voiceprint” of the brand's conversation. Take a look at the full report at http://metaphic.in/nokia/somaReport.php What is the shape and tone of your brand’s voice? Now, we can show you by extracting a "voiceprint" from text data. Metaphic Takes a minute or so. Lacks ux/ui so just click on stuff. Welcome feedback manian@conversation.co.in Grammatically Different.