For the past several years, there has been a substantial amount of rhetoric in the field of Data Management pertaining to empowering the business user through better business analytics.
Numerous vendors, research companies and even trade publications have cited an emerging trend to put the business in control of its data, granting the end user a novel degree of access and autonomy as he or she emerged from the shackles of an IT dominated data landscape.
Developments in business analytics and Business Intelligence were frequently credited as paving the way for the newfound focus on the business, as Data Discovery tools and Cloud-based analytics options abounded, each touting an expedience and “intuitive” ease of use that were heralded as drastic improvements over the analysis of historic data provided by conventional IT.
According to BeyondCore CEO and founder Arijit Sengupta, however, such rhetoric is not always true.
“I actually think that the talk has been all about the business user; the business hasn’t really been the focus. You’re actually pulling a fast one on the business users. What you’re actually telling them is, ‘hey business user, we’re giving you this tool. You go draw the graphs, you go ask questions, you go figure it out.’ That’s not solving their problems. Business users don’t want to become Data Scientists; they want to do their jobs.”
Doing Their Jobs
With BeyondCore’s titular business analytics solution, doing those jobs could become much more accessible, expedient, and even intuitive. Using the product, business users are able to:
- Determine the data characteristics of arbitrary structured and semi-structured data by simply pointing the software at it, without utilizing any code.
- Automate the Data Discovery process courtesy of the solution’s penchant for looking at what Sengupta states are “all possible variable combinations” to determine which facets of a data set are worthy of scrutiny.
- Automate statistical testing to heavily influence modeling by analyzing what Sengupta calls “every statistical test you need for every possible variable combination.”
- Automate the publishing process by creating graphs with commentary that explains the graph’s most salient points and the underlying reasons behind them.
Conversely, business users should also note that:
- BeyondCore was not designed to work on unstructured data, although machine generated, sensor driven Big Data is a form of semi-structured data that BeyondCore handles
- Although the product can be deployed with on-premise appliances or through VMware, its speed and scalability are optimized with Cloud deployments
- The solution has limited utility with small data sets
Automating Data Science
One of the most attractive aspects of BeyondCore is the fact that, as it largely pertains to the business user, the product has effectively automated the functions of Data Scientists. The majority of the statistical work that these professionals do pertaining to data modeling and the various types of analytics—diagnostic, descriptive, predictive and prescriptive—is automated so that individuals with no basic knowledge of statistics can still utilize the benefits of this discipline. Subsequently, end users are able to focus on carrying out their business objectives without the need of difficult to find Data Scientists.
Or, in the case in which an enterprise already employs Data Scientists, they are free to work on more difficult problems and advanced job functions.
“The techniques we are using are exactly what a highly trained Data Scientist would use,” Sengupta revealed. “We have taken exactly what a perfect Data Scientist who had infinite amount of time, did not have a fight with his spouse that day and is not obstructed by anything, does not have any human error, does not have any biases nor political agenda…That’s the Data Scientist we have built [in BeyondCore].”
How It Works
BeyondCore presents the user with all data variables that can be graphed on the Y graph. These need to be numeric; the variables on the X graph do not. Once the user selects a particular variable of focus, the solution analyzes all of the other variables and combinations of variables in the data, and then explains the initial number with its graph and commentary. Thus, end users simply have to select the variable they want analyzed on the Y axis and, with a single click, the solution does the rest. Users are not only able to readily ask questions, but also to gain insight into which questions they should ask.
The automated Data Scientist analytical prowess of BeyondCore is applicable to most industries that rely on data-driven processes. In the healthcare industry, representatives from McKinsey and Company successfully sifted through a million variable computations across 30 million patient records in approximately two hours. In the retail space, Sears Holdings had been analyzing data throughout its entire chain—which spans the United States and parts of Canada—for months. After ingesting its data into BeyondCore, it soon began “finding factors that they hadn’t found after analyzing their chain data for three months—and all it took was five minutes,” Sengupta recalled.
The advantages of deploying BeyondCore as a Software-as-a-Service (SaaS) are plentiful. Hosted by either HP Cloud or Amazon Web Services (AWS), the SaaS version of the solution is much swifter than the on-premise variety since the Cloud allows it to fully take advantage of its parallel computing capabilities. BeyondCore is optimized for business analytics and extremely scalable, which enables it to account for massive Big Data sets much better than it can small amounts of data. Additionally, Cloud deployments require minimal IT involvement to leverage the product, so that end users can begin performing analytics almost immediately.
The overall flexibility of the solution—its propensity to perform all types of analytics in a variety of deployments to meet the particular needs of an organization—underscores its technology-agnostic approach, which in turn emphasizes its commitment to business users in a variety of industries and use cases. “Our approach is focusing the customer’s problem the way they want it solved,” Sengupta commented. “The focus is not on the technology.”
Perhaps the most redeeming quality of BeyondCore is its nearly single-minded commitment to the business user and the achievement of his or her objectives. Numerous facets of the solution reinforce this ideology. It largely automates the elements of Data Science that are most obscure to business representatives including: determining characteristics of unknown structured and semi-structured data, creating data models, and providing statistical analyses of the most relevant data to explicate the circumstances of data behavior. It even goes a step further in some cases by providing explanations of the outcomes of analytics to accompany their graphic representations.
The software is also relatively simple to use with a point-and-click functionality that spurs the aforementioned processes along with those associated with Data Discovery and the identification and exploration of a number of variables relating to a user’s inquiry. Its Cloud deployments take advantage of the speed and scalability needed to analyze Big Data sets, with minimal configuration on the part of IT.
From a wider perspective, then, the solution appears to sufficiently represent the burgeoning use of business analytics in the swiftly expanding list of data-driven processes found across industries and, perhaps, even walks of life. Such a consideration strikes at the very core of BeyondCore’s creation.
“Deeper analytics is really important,” Sengupta proclaimed. “And if it is, people need to have access to it in a ridiculously easy way. If it is not important, then we are all wasting our time anyway. When we think of the internet revolution, it happened with Netscape. It did not happen with the modem. If you think of the PC revolution it happened with the Mac… If we think of word processing and number processing, it became Excel and Word. If it is important, it will become so easy that anyone can do it.”