DataSift has built a reputation as a provider of Twitter, Facebook and Tumblr firehoses — streams of unstructured data from those and dozens of other social sources, which can then be used in applications to track larger user sentiment and other trends. Today, DataSift is ramping up its presence in the big data game with the launch of Vedo, a processing engine that automates some of those functions usually performed by data scientists to make sense of that firehose data.
Co-founder/CTO Nick Halstead tells me that “Vedo” comes from the Italian “I see,” suggested by DataSift’s chief technical architect Lorenzo Alberton, who helped devise the service, its first major product offering on top of its data services.
With DataSift announcing a $42 million round of funding just earlier this month, Vedo underscores the company’s intention to use some of that financing to expand its products and specifically sharpen its target on the enterprise market. Vedo has been a year in the making, he says.
Halstead tells me that the idea behind Vedo is to offer data processing companies (Simply Measured being one example), enterprises, brands, app developers and other customers an easy way of “reading” the data that comes out of DataSift. “We’ve been very good at curating and identifying rules to get to the right data,” Halstead says, “but a lot of that has been really simplistic.”
Now, DataSift will effectively offer customers three ways of tapping into data, by way of Vedo. Social tech app developers, Halstead says, are likely to have in-house data scientists who will be able to use the extensions from Vedo to add machine learning to their existing applications, resulting in faster development. “Similar to how our single API have lowered their costs, this will help them enhance machine learning but with much lower cost,” he says.
Enterprise customers, he continues, are likely to want a bit more structure in how they interact with DataSift data.
For them, the company has created pre-made classifications and taxonomies, which will be available in a library and will continue to be enhanced and improved. Among the examples Halstead showed me there were feeds that, for example, identified what devices and applications people are using to post to Twitter — there are 80,000 sources in all in this one (!) — and a feed that aggregated all airline-related tweets, which can then be classified into different categories like questions, gripes, praise and so on. Yes, all those United and Easyjet complaints, all in one place.
Here’s how one example of that looks. Visualizations of the data, however, would still need to be realised by another provider such as Tableau, for now at least.
View the Original article
0 comments:
Post a Comment