The Data Science Iron Triangle – Modern BI and Machine Learning

Categories: Analytic Database Data Engineering Data Science Machine Learning

The New Iron Triangle

It is cliché to discuss IT/business solutions as people, process, and technology. Some call it the “golden triangle,” but in this blog, we refer to it as the iron triangle. Since the 1960s, technology has disrupted business through the advent of computing and information management. These systems replaced highly manual operations such as record keeping, finance, and reporting. Companies that didn’t adopt technology were soon out of business or in decline. Only 66 companies from 1955’s Fortune 500 exist today. Many people attribute this to technology adoption, which will have a greater impact going forward. Today, computing is taking on even the higher functions of decision and cognitive support. The cognitive computing path is not easy and is full of potholes and challenges.

Most organizations struggle to unlock data science in the enterprise. To that end, Cloudera offers the Data Science Workbench, a collaborative, scalable, and highly extensible platform for data exploration, analysis, modeling, and visualization. It’s powerful features finally get data scientists, analysts, and business teams speaking the same language. Cloudera has simplified the technology, but the people and process are often a source of friction. That friction is what defines the new data science iron triangle. This new iron triangle is made up of data science, IT operations, and business operations. Understanding the data science iron triangle helps organizations adopt and realize value from this new frontier.

The data science iron triangle is different from the golden triangle in that it is very disruptive, complex, and emotional. This is not a discussion about replacing rooms full of filing cabinets and paper receipts with massive refrigerator-sized computers. It is about replacing human cognition and intelligence with algorithms. Until now this was all very much science fiction. Skeptical? You are already experiencing it today through chatbots with your bank, telecom providers, and many online service providers:

As paraphrased from Wikipedia, “Chatbots are programs that are often designed to convincingly simulate how a human would behave as a conversational partner, thereby passing the Turing test. The Turing test is all about how your algorithm simulates human behaviors. A successful Turing test means that your algorithm is very human-like. Chatbots are typically used in dialog systems for various practical purposes, including customer service or information acquisition.”

Why the Data Science Iron Triangle Matters

The three components of the data science iron triangle all have their challenges and strife. Only when organizations understand these challenges will they begin to harmonize and put them to work in a seamless fashion. Below we deconstruct three data science iron triangle dilemmas.

The IT Dilemma: IT is faced with a whole new set of challenges. How do I wrangle in my data science community? How do I support these so-called ‘mad scientists’ and their exotic technologies, algorithms, and mathematics? How do I govern these systems and development processes while also managing quality? More importantly, how can we put them to good use and achieve positive ROI?

The Business Dilemma: Data science is complex and has its own language and perceptions. As a result, friction exists between data science and business communities. It is not only nomenclature and vernacular, but raw fear and emotion. Human cognition has never been challenged by technology and automation, and this scares people. How do we get business to address these trust issues so people still feel like they are in control?

The Data Science Dilemma: Data scientists don’t want to be harnessed by overbearing IT processes and a lack of faith from the business. They don’t care, but they should. Data science without IT and business support will become your organization’s Spruce Goose. Data science needs to succeed in your organization. What is the amalgam – or better yet – what is the glue that can bind business, IT, and data science together?

The Solution

The solution to these data science iron triangle dilemmas is the modern business intelligence platform. The modern BI platform provides a non-intimidating analytic platform for all data – big and small. Being able to scale up and scale down rapidly is key. Solutions like Cloudera Altus give enterprises the ability to perform analytics on big data in the cloud. Leveraging tools like R, Apache Spark, and Python enable data scientists to perform advanced analytics at scale. With the Cloudera Data Science Workbench, analysts can solve high-impact business problems like next best action/offer, churn prediction, and preventative maintenance.

The icing on the cake is the ability to visualize it all seamlessly through business intelligence tools built for Cloudera. Business intelligence is the glue that binds traditional BI and data science together. The goal is to enable business people to operationalize and put data analytics to work. To do this we need a BI tool built for Cloudera. Arcadia Data has created a BI tool for Cloudera that sets a new analytics standard, native to the data lake. It allows you to connect to data, build semantic datasets, visualize, and optimize through a single pane of glass. With Cloudera and Arcadia Enterprise, organizations can break down the data science iron triangle through rapid visualization of data science outputs. You also don’t have to move data, and you can combine the workloads of machine learning and BI together on the same platform. IT ops, business ops, and data science teams require tools that are consumable, friendly, and puts the human in control.

Visit Cloudera’s Solution Gallery to find out more about Arcadia Enterprise or take Arcadia Instant out on a test drive; it’s free, easy to learn, and will not time-out!

by John Thuma, Director of Analytic Solutions, Arcadia Data (@AnalyticsRNA)


Leave a Reply