TextViz Vaidyanath Divya
Why TextViz ? ● Text visualizations of large corpora are hard to comprehend ● Underlying text processing techniques play an important role in making the visual efficient ● Its nice to have a platform to visualize the text in various ways
Project Goal Create a one stop efficient text visualization platform to: ● Gather basic text statistics ● Understand the text corpus ● Help data analysts get an idea of the text to further apply models - Text mining
Components Data input from user (Html/CSS/JS) ● Single text document ● Multiple documents Text processing (Python/Flask) ● Basic text statistics ● Summarize text ● Similarity between text documents ● Topic detection
Components continued . . . Visual element (Html/CSS/JS) ● Word Clouds - With contextual component ● Occurrence of words in a text ● Cluster of topics (Bubble charts)
Existing techniques/ Prior work ● Cui, Weiwei, et al. "Context preserving dynamic word cloud visualization." Pacific Visualization Symposium (PacificVis), 2010 IEEE. IEEE, 2010. ● Miller, Nancy E., et al. "TOPIC ISLANDS< sup> TM</sup>-a wavelet- based text visualization system." Visualization'98. Proceedings. IEEE, 1998. ● Pan, Shimei, et al. "Optimizing temporal topic segmentation for intelligent text visualization." Proceedings of the 2013 international conference on Intelligent user interfaces. ACM, 2013. ● http://www.chrisharrison.net/index.php/Visualizations/BibleViz ● https://gramener.com/mahabharatha/
Demo
Milestones Date Task Deliverable Nov 18 2013 Research / Initial design storyboard / basic framework Nov 21 2013 Complete research and build Initial backend algorithm in backend text processing python algorithm Nov 25 2013 Experiment with various visual Initial frontend layout displays Nov 30 2013 Sync frontend and backend Initial complete tool Dec 2 2013 Quality testing and Final model/tool modifications Dec 6 2013 Complete writeup Project complete
Thank you Any Questions ?
Recommend
More recommend