Closing the Gap – Handling mixed data to investigate emotions

What’s the challenge? One problem has been that traditional approaches to data investigations have been lacking the ability to handle mixed data effectively. Enterprise systems allow to search across structured and unstructured data, but just to reveal “results” in the form of – for example – blue links. And Business Intelligence systems on the other hand are great for seeing aggregates but lack the ability to search and investigate down to the single last record.

Within the context of the MixedEmotion project we’re interested in investigating the “emotions” in texts. At the core are algorithms, which extract measurements of “Emotional states” from text. The harvested emotional data can then be measured, averaged and acted upon.

Need to search, slice & dice results

Within the context of identifying emotions and making sense out of them, analytics plays a fundamental role. For example algorithms can be often noisy, when interpreting human communications. Therefore it can be helpful to see the “big picture” which often averages the results across large samples. Another important aspect is to be able to search, slice, dice, see results divided and grouped by all sort of possible variables in order to reveal effects and correlation – and ultimately inform decisions. This again is analytics.

 

Search Structured versus unstructured Text

Closing the gap between traditional text search and business intelligence systems

 

Kibi is one approach to close the gap. Kibi is an Open Source platform for Data Intelligence based on the search engine Elasticsearch. While Kibi looks may seem similar to those of business intelligence tools, using Kibi feels different, because it is, at heart, based on a quite different technology.

Kibi search

Kibi allows to both aggregate and being able to investigate down to the single last record

Kibi is a “friendly fork” of Kibana that brings the same snappy awesome search & analytics to data intelligence use cases. Data Intelligence scenarios revolve around relationally connected data which is typically more complex than the one handled by Elasticsearch and Kibana out of the box.

From faceted browsing to “relational faceted browsing”

With Kibi one can, at a press of a button, filter across dashboards and across entities at the same speed as Kibana. Instead of “faceted browsing” – the usual way one drills down in ecommerce shopping experiences – one gets cross entity “relational faceted browsing”. Ok, easier shown than explained so please do take alook at this video to see what i mean 🙂


And the same search analytics engine is ready for more. In fact here are some more use cases where Kibi can be pretty useful:

  • Business Intelligence with Text Analytics: Elasticsearch is a search engine at heart, so mixing filtering data using advanced search queries is a breeze. What are the most purchased products by customers that during any email or support interactions have negatively mentioned the name of a competitor in the past quarter?
  • Law Enforcement: monitor hotspots of negative feelings about specific topics.
  • Social / Smart TV: allow emotions to be used as part of a recommender system and allow editors to understand emotional reactions to specific topics, merging them from social network streams and from the news themselves.
    Brand Reputation Management: monitor the emotions of a brand overtime doing the analytics of the information streams from social networks and online forums

… just to name a few.

Pushing “Search” even further

This is just the beginning. We have a very exciting roadmap ahead. Those familiar with the most advanced new features of search engines (e.g., the new streaming aggregations in Solr 5), know that “search engines” have quite some exciting aces up their sleeves and roll out and there is no sign of these advances “slowing down”. At SIREn Solutions we are sure that these advances will lead to Data Intelligence capabilities which will be not only powerful but also much easier, thus catering a wider user base in the enterprise.

Search engines have always been the the easiest and most natural way to dig into data; their snappiness and the inherent feeling of fresh data being immediately available make them a pleasure to use. And today they can return you much more than just blue links.

Kibi homepage is http://siren.solutions/kibi

Open Source
Kibi is free and open source. We offer commercial support and 24/7 assistance and you might be interested in the enterprise features we will be releasing soon, but feel free to experiment with our demo distribution for any use, right away.

About Giovanni Tummarello, Ph.D: Giovanni Tummarello is CEO of SIREn Solutions, a company cofounded with Renaud Delbru, specialised in data intelligence scenarios and big data search scenarios also involving extensions to Solr and Elasticsearch.

Note: Data is from the Crunchbase 2013 Snapshot ©, Elasticsearch and Kibana are trademarks of Elasticsearch BV, registered in the U.S. and in other countries

Photo by Lindley Yan (Flickr)