OneFold believes that the expertise of a data scientist should be used for solving the most complex of analysis and not for the more day-to-day aspects of data extraction. The company uses a plug and play DHQL platform for automated data extraction, transformation, querying and reporting. Navneet Aron and Uday Sandhar, co-founders of OneFold, teamed up to answer our questions about this intriguing technology.
insideAI News: OneFold seems to be taking a whole new approach to Big Data. What is this approach and how is it indeed different?
Navneet and Uday: Some competitors believe that the data scientist can be eliminated. Others are tools to exclusively make the data scientist’s job easier. We are the first company that is taking a practical approach to dividing the analytics problem in big data by the 80/20 rule: 80 percent of analytics will be directly available to the product executives without data scientists, and 20 percent of the analytics work will still be done by data scientists.
OneFold solves a fundamental tradeoff between having to use expensive resources to generate deep insights, or getting only basic analytics using off-the-shelf tools. OneFold delivers deep analytical insights without the need for expensive data scientists by flipping the traditional ”Data Light, Query Heavy” model of BI tools that are not used to unlimited storage at virtually no cost, into a Data Heavy, Query Light model.
It’s important to note that OneFold is not a data collection tool. OneFold is not based on the dated architecture utilized today by most of the analytics tools which require heavy dependency on specialized data scientist resources.
insideAI News: As this technology emerges, who will be your customers?
Navneet and Uday: Organizations with mobile and wearable apps that generate enormous amounts of data are OneFold’s target market. We are focused on Product Managers, User Experience Managers and Marketing organizations so they can optimize mobile app revenues (projected to reach $200B yearly run rate by 2017) in m-commerce, travel, lifestyle, financial and gaming verticals. OneFold co-exists with these players by focusing on empowering the App Owner directly with a killer product they love.
insideAI News: How are you leveraging massively parallel cloud computing?
Navneet and Uday: OneFold uses a new architecture based on our cloud-based “DHQL” technology that leverages recent inflections in the price of storage and massively parallel cloud computing to deliver deep analytical insights. To leverage this inflection, OneFold created a plug and play insights platform which pulls all historical and incremental event stream data from any JSON-based data source. It denormalizes the data into a very broad schema, and stores in its Google Cloud based columnar database, which has massively parallel computing available. Then, OneFold’s backend generates a number of transformations that are visualized on our intuitive tablet app and can be easily manipulated by product managers to get deep user insights.
insideAI News: How easy is it for an organization to deploy the platform? What do they need to get started?
Navneet and Uday: To get started, all the organization needs is to provide OneFold with its raw storage of JASON data. After that, OneFold delivers the user the user experience on tablets as described below. We’ve created an insanely intuitive tablet app that gives Product Managers all user pathflows visualizing every possible user funnel — no more hypothesizing user behavior — that turns the query-able views into real time, slice and dice visualizations.
PMs get views by any collected or computed event or event attribute. All this can be done without requiring a data scientist to do extraction, transformation, querying and reporting, allowing data scientists to focus on extremely complex analyses and models that require their true expertise.
insideAI News: This is pretty bleeding edge stuff. What tricks do you have up your sleeve that we may see down the road?
Navneet and Uday: OneFold’s DHQL platform combines columnar databases (pioneered commercially by Dr. Michael Stonebraker founder of Vertica, and Jerry Held ex-Chairman of Vertica and now advisor to OneFold), and elastic cloud processing, with in-memory query techniques developed at AMPLab (led by Dr. Michael Franklin of UC Berkeley, now an advisor to OneFold) into an integrated architecture optimized for real-time analytical visualization of big data.