Visualization is not business intelligence

Fancy charts and graphs portray a very misleading view of business intelligence especially when the underlying data sets, and the compute required to make sense of these data sets is either entirely missing or highly disorganized.

Sep 24, 2021 11:09 IST India Infoline News Service

Business intelligence has long been touted as the key benefit of automation and digital transformation. As organizations seek the make sense of the plethora of data resulting from rapid digitization and automation, business intelligence continues to be elusive. Most of this can be attributed to one factor – visualization tools have been positioned and sold as business intelligence systems. The first step to ensuring that your data talks to you is to dispel the notion that visualization tools are the organization’s business intelligence layer. Fancy charts and graphs portray a very misleading view of business intelligence especially when the underlying data sets, and the compute required to make sense of these data sets is either entirely missing or highly disorganized.

There is no business intelligence without data management

With vast amounts of structured and unstructured data being collected each day, automatically tabulating this data, building relationships between data sets and dynamically modify data models is critical. It is equally important to reduce data latency and stream data from various sources without increasing the number of trips required to cleanse, transform and prepare data. The traditional method of ETL (Extract, Transform, Load) pipelines placed between source-to- target systems is simply not conducive to low latency high speed decision-making. For business intelligence to work, all these aspects need to be addressed through dynamic data management, pre-packaged data streaming, API management and on-the-fly data transformation in the business intelligence platform.

Compute is the heart of business intelligence

From raw data to organized data to business intelligence, a key component ignored is compute. Compute has for long been splattered over ETL tools, using disparate calculation engines, scripting computation logics within a visualization tool, and writing external scripts to somehow convert data into meaningful information. As a result, this makes compute lose transparency and control, and dramatically increases the cost of maintenance. A unified compute layer that can ingest data, transform data and perform complex statistical and ML operations is critical for business intelligence. What is equally critical is the ability to visualize and see the informational output at each stage of computation to enable corrections, enhancements and adjustments on an ongoing basis.

Data management and compute is fragmented and laborious

Business intelligence today is still an expensive, high latency and difficult-to-scale proposition requiring a combination of largely redundant ETL technologies and a plethora of cloud services stitched together in an awkward manner, and in a very use case specific format. Stitching together cloud services such as event response for streaming data, streaming pipelines, interim storage, in-memory storage, compute algorithms, compute management services, queuing services, ML models, visualization, information distribution services and orchestration tools makes for a very bloated environment requiring levels of expertise that are not expected to be common for most organizations. Fixing business intelligence requires pre-architected, integrated and easy-to-deploy business intelligence systems or platforms that not only promise visualization but also empower a do-it-yourself approach for data management and compute.

Integrated plug-and-play business intelligence is the only effective way to operate today

A new generation of business intelligence platforms is all set to replace the traditional visualization tools backed by bits and pieces of data management and compute technologies. These new generation of business intelligence platforms provide an out-of-the-box plug-and-play capability with in-built and integrated data management and compute capabilities. It is time for organizations to re-think the charade that visualization is business intelligence. Heavy investments in failed data warehousing projects and the high cost of building and maintain dashboards are a thing of the past. Business intelligence in today’s context has to mean integrated data management, compute and visualization for high-speed deployment and low-latency decision-making.

It is time to replace current visualization tools and high-cost data warehousing projects with high-speed, low cost and pre-architected true business intelligence platforms. The future of business intelligence is simple to use, quick to deliver and meaningful to end-users.

The author of this article is Mr. Muzammil Patel, Global Head Strategy & Corporate finance, Acies

The views and opinions expressed are not of IIFL Securities,

Related Story

Open Free Demat Account (Rs699)
Open ZERO Brokerage Demat Account

  • 0

    Delivery Brokerage for Lifetime

  • 20

    Per order for Intraday, F&O, Currency & Commodity