SRK Does Not Require Data Warehouse

SRK innovative and ease-of-use importer

A data warehouse does not have to be built for SRK, because data modeling is not a prerequisite of using SRK. Out of the box, SRK can easily import data reside in CSV/Excel files that are extracted from your operational systems and/or importing data from mainstream relational databases (Oracle, MS SQL Server, etc.). Often times, business users have been working with those spreadsheets already, in that case they can easily use SRK in lieu of Excel to come up with the analysis. SRK Importer – that allows business users to quickly setup data import with minimal IT involvement (if not at all), even data cleansing and transformation can be done without requiring any programming.

SRK is usable immediately after bringing source data into the platform. The idea is that analysis (and the data to support it) is typically grown organically, meaning users will be building analysis naturally based on their understanding of the data and their analytical requirement, other than being forced to abstract their business requirements and design the underlying data models upfront. And this stands true even when building multi-dimensional analysis – by incorporating Big Data technologies such as Hadoop, SRK is able to dynamically generate the multi-dimensional data set (equivalence to OLAP cube) on the fly, which enables business users to quickly create drilldown-enabled multi-dimensional analysis via an easy to use, drag-and-drop user interface. As a self-service BI tool built from the ground up with business users in mind, the design of SRK facilitates that.

Once the initial set of analyses is created, the use of SRK would help identify new analytical requirement and will further spawn new/ad hoc analysis around the existing analysis. So, SRK users are able to create sophisticated and inter-dependent analyses without even worrying about data modeling. One may question that, in order to support sophisticated analysis, there have to be some sort of “structure in the analysis”. That is true, but in SRK that is accomplished in an entirely different, and greatly simplified way. The business user just need to identify matching fields across multiple data sources to bring them together in one analysis, as well as which fields they want to aggregate and/or group by in order to build a multi-dimensional analysis. But this is not too different from, say, when they use spreadsheet to do so, only more intuitively. And by recursively iterating this simple process, every time with more data brought in, business users eventually creating some sort of “data relationships” (which can be inspected from the Dependency UI of SRK), but the key point is they are doing this spontaneously. This different paradigm SRK has taken allows business users to focus on the business questions they wanted to get answer for, rather than been forced to think from pure technical perspective and build the underlying abstract data models first.