Sonar Solo visualizes trends in the Twitter fire hose
Bottlenose's free Web service dynamically exposes related concepts or trends for a query in real time, and allows users to continuously drill down via its visual interface.
Bottlenose is offering a deeper and visual look into what's trending in the global Twitter fire hose with Sonar Solo.
"Sonar Solo is the only live visualization of breaking trends in the global, social fire hose. This is not your daddy's static tag cloud. It shows real-time changes and trends right when they pop via a massive amount of data mining and natural language processing in the cloud," Bottlenose CEO Nova Spivack said Wednesday at the Pivot Conference in New York.
Unlike traditional Twitter or social search tools, Sonar Solo dynamically exposes related concepts or trends for a query in near real time, and allows users to continuously drill down into trends via its visual interface, according to Bottlenose. The type size of text shows the relative volume of citations, and an orange color indicates that a term or topic is gaining tweet velocity. Green and red dots indicate positive and negative sentiment.
"We have developed the most powerful sentiment analysis system that combines technology from Lexalytics with our own algorithms for detecting emotion and sentiment bursts," Spivack said. "The combined system has been tested at 80 percent accuracy, which is better than humans typically."
Sonar Solo is a free version of Bottlenose's enterprise product that is used by companies including Pepsi, General Motors, and Warner Bros. It analyzes just 1 percent of the tweets for a given search, compared with 100 percent for the Bottlenose Nerve Center enterprise edition, which can cost upwards of $100,000.
Pepsi, for example, uses the paid version to track millions of relevant topics, hashtags, and people, and mines more than 2 billion time series continuously to capture trends and sentiment around Beyonce, who the company sponsors.
"Sonar Solo delivers a statistically valid result. On 500 million messages, 1 percent is 5 million messages being analyzed continuously," Spivack said. "For any high-volume topics there are usually thousands of messages flowing in, and this works great. We also pull in the most relevant tweets in a window on the right of the screen. It provides another way to see what's trending."
Bottlenose plans to offer other free tools and "experiments" from its labs, Spivack said. The company has 22 people and 22 total patents pending for its social data trend analysis technology.