kryndex / sententree

A novel text visualization technique

SentenTree

SentenTree is a novel text visualization technique for summarizing a collection of social media text, i.e. take thousands or more Tweets and summarize what the Tweets are about. The aim of this project was to create a visualization that is cheap to compute but represent the connected thoughts in the words.

See DEMO

Author

• Mengdie Hu / @mengdieh
• Krist Wongsuphasawat / @kristw

Publication

Mengdie Hu, Krist Wongsuphasawat and John Stasko. Visualizing Social Media Content with SentenTree, in IEEE Transactions on Visualization and Computer Graphics 2016.

Installation

npm install sententree

Example usage

<div id="vis"></div>

d3.tsv('data/goal.tsv', (error, data) => {
  // data format is [{ id, text, count }]

  const model = new SentenTreeBuilder()
    .buildModel(data);

  new SentenTreeVis('#vis')
    // change the number to limit number of output
    .data(model.getRenderedGraphs(3))
    .on('nodeClick', node => {
      console.log('node', node);
    });
});

For developers

Install dependencies via npm or yarn

$ npm install

Then run local instance via

$ npm run start

License

Copyright 2014 Twitter, Inc. Licensed under the Apache License Version 2.0