Real-time social analytics platform Topsy, which earlier this month debuted Twindex to provide insight into Twitterati sentiment on the presidential candidates, today unveils Topsy Pro Analytics. It delivers in-depth metrics based on the Twitter firehose via API to the general public. Previously, the company had API access for some metrics in a machine-to-machine interface, but nothing near the full interactivity nor access to all the measurements that are propagated into the new user interface.
Topsy’s technology was created to ingest huge amounts of authored content, with Twitter as its primary data source — all 400 million tweets a day, with an index that goes back multiple years. Topsy also does a full public scrape of Google Plus and indexes that data. It offers its own sentiment classification and dictionary scheme tuned for tweets, takes every link published in tweets and unpacks them to their native states to produce measurements around them, provides a geoinference model to see where people are communicating from (to the country level today but soon to city and state level), and also can deliver an influence and author graph.
“The new UI lets people take advantage of all the outputs from that processing,” says Topsy chief revenue officer Eddie Smith. From the Topsy Pro Analytics dashboard users can search terms to analyze top posts (those most relevant to the search query), top links, and volume metrics, and also have the service propagate up new terms related to those they are analyzing (perhaps a new way people are referencing a product) to grow their data points and expand their analysis. “There is no other place on the planet right now where you can take any keyword or term and look at real-time and multi-year historical results for that term using the full corpus of Twitter,” Smith says.
Use cases he proposes for such analysis range from marketing to publishing – any industry doing content analysis could find value in understanding and quantifying those signals from Twitter for their own or competitive domains, and even dive deeper with a click, for instance, to drill in and see what drove a spike. “So it’s not just metrics but you can go back in time to see significant tweets that drove that,” he says.
Besides keyword analysis, part of the service’s processing is unpacking every single link communicated in tweets. “In the UI that gives very powerful ways to do domain and link analysis using Twitter as a signal to understand what content people are paying attention to,” says Smith. “So Google which scours the whole web is a huge corpus of data, but you don’t really understand what is important or what people are paying attention to. Our signal to understand what content is most important uses various measurements to quantify the extent of that attention.” Although it doesn’t index Facebook data, as that service doesn’t offer a full public API, the links people communicate over Twitter provide the service with a lot of signal about what people are paying attention to there.
Trending metrics focus on momentum, acceleration and peak. So, users can discover, from the weight placed on more recent activity for momentum, when something is taking off now, or when something topped out in interest. “For content publishers and those trying to understand what links on the web are starting to trend by any domain, this is powerful analysis. It can give publishers a sense of where to look next or what stories to follow up on,” Smith says. Agencies, he notes, use such analysis to understand what content is trending across a group of domains, say, those related to health and beauty, so that they can play to that popularity in the content they’re producing.
Providing insight into where tweets hail from lets users see all mentions around the globe of the keyword they are exploring, identifying what country each mention is from and providing distribution by country of the number of mentions and more importantly the relative concentration of mentions by country for that term. “Just the number of mentions tends to skew to population centers,” Smith notes.
Topsy Pro Analytics also will be available at what Smith says will be “disruptive public pricing.” He won’t provide specific costs but says that all social analytics solutions up to now take the method of grabbing data, analyzing and presenting it, which has a variable cost. Topsy, however, “already has all the data, so we charge a flat monthly fee for all analysis you can do in real-time as far back in time for as many keywords as you want.”
- MESA Announces New White Paper: The Role of Semantic Models in Smarter Industrial Operations
- Automatic Hashtags & Machine Learning: The New Google+
- Cambridge Semantics Wins SIIA Software CODiE Award
- INTELTEQ Named a Cool Vendor in Information Governance and MDM