Q1. What is the difference between business intelligence and data analytics?
We are seeing business intelligence being used as a broad term encompassing various elements a business needs to get insights to make data-driven decisions. Traditionally data analytics implied algorithmic modeling and applying data science techniques for predictive and prescriptive analytics, in addition to descriptive analytics provided by most BI tools. But without getting caught up in the semantics of the terms, what is important is that business users are looking for tools that can help them make informed decisions, and this is broadly grouped under business intelligence initiatives.
Q2. How do you eliminate the need for top-down data modeling and ETL when performing data analytics?
There is always going to be a need for making sure that data is clean, validated and governed to avoid “garbage-in garbage-out” scenarios. However the business user should not have to be burdened with having to do the heavy lifting and modeling. We see a number of companies building semantic models and providing guide rails for business users to navigate the data available in their big data environments This does not take away their ability to explore, but actually makes the experience much more productive.
Q3. What is self-service business intelligence and for what it is relevant?
A business user should be able to get insights from data without having to rely on programmers to write complex SQL queries or build reports. They need the ability to explore, slice and dice, and navigate through the data, as well as drill down to the lowest levels of granularity to identify patterns or anomalies. But running unconstrained SQL queries on massive amounts of data can bring a cluster to its knees. That is why we see companies defining and building BI Consumption Layers on their big data lakes. This provides the semantic definitions, and it also includes pre-built aggregations and computations so that at query time the business user can get an interactive response. This transforms the self-service experience for a business user to interactively do analysis on massive data, without having to submit jobs and wait for reports to be generated the next day.
Q4. If business users have tools to analyze business data directly, without the need for data architects and data scientists, who can guarantee that the insights they obtained are correct and relevant?
This is why it is important to make sure that data has been cleansed, validated and has the governance and security structures in place. Beyond that, the semantic definitions need to be clearly defined. Once these “guide rails” are put in place, a business user can freely explore and analyze the data and have a high level of confidence in it. At Kyvos, our focus is on making this experience the most rewarding for business users, so that they can explore and get insights from the data in a self-service and interactive way. The results and benefits our customers have seen have been transformative for their businesses.
Ajay Anand is the Vice President of Products & Marketing at Kyvos Insights. His association with Hadoop goes back to 2007, when he was Director of Product Management at Yahoo and led their initial Hadoop projects and releases, after which he founded Datameer. Ajay also held product management and market development roles at SGI and Sun. Ajay earned an M.B.A. and an M.S. in computer engineering from the University of Texas at Austin, and a BSEE from the Indian Institute of Technology.