When OpenText launched Election Tracker ’16, we received several encouraging and positive responses regarding how easy it is to compare stats about their favorite Presidential candidate using the interactive visualization and intelligent analysis.
And without fail, the next question was, “How does it work and how could it help my business?” Powered by Release 16, Election Tracker is a great example of unstructured data analysis in action. It is able to showcase the power and importance of unstructured data analysis in a relatable way. In fact, we feel the Election Tracker addresses the dilemma of unstructured data in three distinct ways.
1) Intelligent Analysis
Making sense of unstructured data is a high concern for digital organizations. Perhaps it’s trying to understand Google’s Page Rank algorithm, finding sentiment in the body of an email or website, or perhaps scanning digital health records for trends. It is also important for businesses that need to organize and govern all data within an enterprise.
Companies are not shy about throwing money at the problem. The global business intelligence market saw revenue of nearly $6 billion in 2015. That number is expected to grow to $17 billion at a CAGR of 10.38 percent between now and 2020, according to market research firm Technavio. Much of the investment is expected to come in the form of data analysis and cloud implementation.
The secret sauce is our content analytics tool, OpenText InfoFusion. Using natural language processing technology or text mining engine, the software tackles the core of unstructured data by extracting the most relevant linguistic nouns from semi-structured or unstructured textual content. The extraction is based on controlled vocabularies such as names, places, organization labels, product nomenclature, facility locations, employee directories, and even your business jargon.
Want to build and benefit from your own data analytics?
Join OpenText Analytics in San Francisco on Thursday, Nov. 10, for a complimentary one-day interactive, hands-on analytics workshop that dives deep into enterprise-class tools for designing, deploying, and displaying compelling data visualizations.
Click here for more information and to sign up.
The InfoFusion engine is able to automatically categorize content based on a file plan, hierarchical taxonomy or classification tree. This automatically creates a summary combining the most significant phrases and paragraphs. It can also show related documents. This ability to relate documents is based on semantics—asking the engine to give you a document that has the same keywords, key phrases, topics and entities. The engine can also detect the ways that key words and phrases are used and correlate them to known indicators of whether a document is dryly factual or conveying emotion about a topic, and whether that emotion is positive or negative ─that is, its sentiment.
2) Interactive Visualization
All the data in the world means nothing without some way to visually represent the context. Most pure-play “text analytics” solutions on the market today stop short of actual analysis. They are limited to translating free text to entities and taxonomies, leaving the actual visualization and analysis for the customer to figure out using other technologies. The technology powering the Election Tracker overcomes this important dilemma by converting the data into a visual representation that helps with content analysis.
Once the Election Tracker mines raw text from the scores of major news sites around the world, it then uses OpenText Content Analytics to process the content. This determines sentiment and extracts people, places, and topics following standard or custom taxonomies providing the meta-data necessary to conduct an analysis. The tracker determines the objectivity or subjectivity of content and the tone: positive, negative or neutral.
3) Scalable and Embeddable
While we designed the Election Tracker to automatically crawl the web for election-focused articles, the technology behind the scenes can access and harvest data from any unstructured source. This includes social sites like Twitter, Facebook, and LinkedIn; email; multimedia message service (MMS); document archives like PDFs; RSS feeds; and blogs.
Additionally, these sources can be combined with structured data to provide extremely valuable context—such as combining brand social sentiment from Twitter with product launch campaign results from a customer relationship management source, giving unparalleled insight to the success of a launch process. Overcoming the problems of scale can help ease fears about needing to add more data sources in the future. Its ability to be embedded allows companies to use their own branding and serve their customers in a format that is comfortable to the end user.
See what all the buzz is about by visiting Election Tracker ’16 at: Electiontracker.us
For more on the technology behind the Election Tracker ’16, here is a 20-minute election tracker demo.
Also, visit our blog to discover the importance of designing for unstructured analysis of data.