Twitter Visualization: Semtech Data Feb 2013
1 like591 views
Visualization tool used to see if ai-one's biologically inspired computing can discern meaningful associations in the mess of tweets from a technical conference. This capability serves as the foundation for building intelligent agents and other applications allowing human interpretation of large data sets.
Twitter Visualization: Semtech Data Feb 2013
- 1. Outline of discussion Topic-Mapper: ai-one for Text Twitter Use Case Semtech Data - Title Schema.org example Tom Marsh February 2013 ai-one biologically inspired intelligence
- 2. Can ai-one make sense out of twitter data? Source data is a file of tweets #semtech from days during the conference- Total file is approx 7800 lines of tweets, original data shown at right. Messy original and clean data are used in this case cleaning included removing time and url info
- 3. Assoc Tree for schema.org net is from semtech tweets Circle indicates area zoomed on next slides Schema.org was mentioned over 120 times and was a subject of discussion at the conference as industry was not pleased at the announcement by Google, Microsoft and Yahoo of their own standard
- 4. Words of Highest Relevance closed-doors Sticking Supportive Countless Bitch Tagging Responses Pre-orders Trunk Tap Jewish Holiday LinkedData Etc. These words were checked in source data for relevance to schema.org
- 5. This image shows with just clean data
- 6. This image versus first slide shows increased quality of semantic after importing messy tweets on top of clean tweets Demonstrating simple programmatic enhancement without losing original associations
- 7. This image shows quality of semantic after importing messy tweets Displays all relevant terms when zoomed in to show word bitch
- 8. This image shows increased quality of semantic after importing clean tweets on top of messy tweets Note fewer terms of greater relevance (14 vs 18+)
- 9. Did the word make sense? The word bitch was used in only one tweet and accurately paraphrased the conversation 7:35 nopiedra <http://www.twitter.com/nopiedra> #linkeddata RT @paul_houle <http://www.twitter.com/paul_houle>: was a bit depressed that people at #semtech would rather bitch about schema.org than swap war stories
- 10. Your brain on twitter.. Need to understand how your audience sees twitter? Our new visualization tool allow you to filter, tag, and act on twitter streams. NathanApp was announced in Zurich February 27th and Baja Mexico February 28th. Now developers have the tool they need to make big data personal.
- 11. Tom Marsh, COO ai-one inc. 5711 La Jolla Blvd., Bird Rock La Jolla, CA 92037 phone: +18585310674 Twitter: @tom_semantic www.ai-one.com www.ai-braindocs.com 息 ai-one inc. 2013