Researchers Find Business Value in Tools for Sentiment Analysis

Content searchers are gaining new insight by measuring responses to news and social media

Editor’s Note: InfoDesk provides the services described in this article – including social media monitoring and sentiment analysis – through our curated content solutions or integrated into more extensive projects. The free offerings mentioned here add even more business value when coordinated with curated content and editorial insight.

Information professionals are constantly trying to find added value tools to bring insights to vast amounts of data. A whole new generation of such services is now becoming available, providing sentiment analysis, positive / negative “scores” and geo-located heat mapping of news and comments. Online tools for sentiment analysis have been proliferating, allowing researchers and analysts to gain deep insight into news stories and social media commentary. Assorted sentiment tools are profiled below, offering a broad array of analysis capabilities.

Sentiment Tools for News and Social Media

IBM Watson has a news explorer tool which can provide heat maps and sentiment analysis. Simply browse the top news or type in a keyword concept, such as “baseball,” and the tool will immediately show recent news on the topic, a word cloud of top keyword connections, a geographic map of top locations of the news and assorted other value-added data. When individual titles of news items are listed, the lighter color font corresponds to articles that are more positive in tone. The topic word cloud also allows the researcher to know what words are being used the most in connection with the given topic in order to understand what the current mood is about that given subject.

The Bing News extension offers an add-on tool that provides researchers with ways to understand themes, related topics and overall sentiment of a given keyword / brand / company. The service is based on machine learning technology using Azure software. A subscription to Azure is needed in order to fully use the tool. The technology organizes the news in groups and themes. A user can gauge what news items are the most positive or most negative regarding a given keyword / concept / brand. A trending area of the tool allows the user to analyze what words/topics are most connected to a given keyword / concept / brand.

Many people are aware of sentiment descriptors like “Bullish” or “Buy” or “Sell” for stocks, but the Sentifi website offers expanded capabilities and analysis for sentiment on publicly traded companies. The tool analyzes up to 90 days’ worth of news, Tweets and blog content in order to evaluate positive and negative events that may affect how a given stock price has fluctuated. Companies, currencies, stock indices, commodities and executives can be searched in this database in order to see the latest sentiment. For example, IBM’s record will bring up a chart that has a comparative line graph of stock price versus various content “Voices.” Additionally, the recent major events that have happened to the company are succinctly listed, noting whether they are a potential risk to the company. Overt risks in the news are separately listed in an “Alerts” area of the record. A listing of some of the recent Tweets, news and blog postings driving the sentiment ratings is also included.
Regarding social media analysis – also called “”social listening”” – 30db is a search tool that provides positive / negative Twitter sentiment analysis. Type in a topic or name / brand, such as “Starbucks” and the tool will show an immediate percentage ratio of positive / negative tone of Twitter postings. Additionally, top keyword descriptors used when discussing Starbucks are also profiled in the record, allowing the researcher to understand what accounts for the positive / negative score.

The Future of News Sentiment Analysis

Added value metrics and analysis of news can be a most welcome addition to a researcher’s tools. Many of these sentiment resources continuously release updates and improved search features. The data that these tools may bring to a researcher may enable connections and insights that may have otherwise not been found.