Loading...
You are here:  Home  >  Data Education  >  Big Data News, Articles, & Education  >  Big Data Blogs  >  Current Article

How Cognitive Data Science Can Cure the Talent Shortage

By   /  May 13, 2016  /  1 Comment

Click here to learn more about author Sundeep Sanghavi.

The Internet of Things and its interconnected approach has led to an overwhelming explosion of big data. About 6.4 billion connected devices are already in use today, and Gartner expects that number to reach 20.8 billion by 2020. This growth is big business; it’s estimated that IoT-enabled data monetization will become a $11.1 trillion industry by 2025.

These astonishing numbers have led to greater demand for high-priced data scientists, experts tasked with digging deeply into massive caches of data to provide analysis that fuels a company’s actions. However, there’s an ongoing shortage of data scientists in the world, and supply is unlikely to meet demand anytime soon.

But even if data scientists grew on trees and came with reasonable price tags, they still wouldn’t be ideal when it comes to swiftly and actionably analyzing data. Dealing with the staggering demand is not a human scaling issue, and throwing more bodies at your mountains of data is not the solution.

The big data explosion must be addressed through a machine-first approach. By embracing automation in data science and machine learning, enterprises can easily sidestep the data scientist shortage to achievegreaterspeed, scale, and repeatability.

The Human Burden

I’m sure you’ve heard about the many companies participating in Kaggle competitions, which demand advanced technical skills in data science, statistics, and machine learning. Well, a rising number of enterprises are also focusing on bridging the data science skills gap by creating citizen data scientists.

Assigning expensive, specialized human capital to data analysis often entails longer execution cycles and higher operational costs, especially considering the number of tools and technologies data scientists need in order to gain holistic views. Data scientists should spend their time exploring the new possibilities of big data; they shouldn’t be spinning their wheels by performing tasks that could be automated.

As a workaround, companies often hire consultants, which frames data as a fixed-term project and neglects its role as a crucial element of day-to-day decision-making. Data analysis should never be limited by a start and end date, which is essentially the cornerstone of consulting.

The true solution companies should embrace involves democratizing data science through machine-based automation — which enables all business and data analysts to become citizen data scientists.

Redefining Data Science

Automation in data science and machine learning represents a quantum leap in the big data era, offering the possibility of algorithms that are not just intuitive, but also able to iteratively learn without manual intervention. These machines can identify, select, and apply complex statistical analyses and recommend the best algorithms for each strategic initiative. With machines doing the grunt work, citizen data scientists don’t need in-depth understandings of algorithms and coding parameters; the automation takes care of that.

This approach is called “cognitive data science,” and it offers numerous benefits. To start, it empowers enterprises to build and execute data initiatives within hours — not months. And further, its algorithms are both repeatable and scalable. They can be incorporated into a company’s workflow, freeing up data scientists to focus on the resulting insights and explore new ways to ingrain data deeper into the day-to-day. MIT scientists confirmed in a study that automation in data science and machine learning represents the wave of big data’s future.

The only way around the data scientist crunch is to empower the broader community and ensure that traditional data scientists focus on solving real-world problems rather than completing manual tasks. Companies should take these four crucial steps to implement this new data methodology:

  1. Find the Problem. Ad-hoc data queries might offer insight on the past, but they seldom help devise strategies. Instead, businesses should focus on using data to solve ongoing problems. Don’t ask why customer churn increased in January; ask what factors caused that churn and how your company can prevent it from happening again. The insight from this query will be far more useful when planning future company growth.
  1. Find the Collaborators. A data product cannot be built by a single engineer; it requires the collaboration of multiple teams with different areas of expertise. Data engineers, data scientists, and business analysts all have their parts to play, and they need to work together from the beginning to ensure the product meets company needs. Remember: There’s no “I” in data.
  1. Find the Platform. Many of the machine learning and business intelligence APIs on the market today are focused on solving specific problems. They need to be combined manually to reap the best insights, and this approach often results in a cumbersome product.

Instead, companies should look for a single platform that offers an end-to-end product. The most effective option will require minimal human intervention. For instance, Amazon’s recommendation system doesn’t require humans to monitor customer browsing and purchasing decisions; it’s all done automatically. Your platform should incorporate that same approach.

  1. Find the Interface. If a data product is difficult for end users to interact with, they will struggle to make the best use of its insights. Natural-Language Question Answering interfaces, such as those used in Siri and Google Now, make the whole process simpler. They allow every person in an organization to interact with the system. Further, they allow the automatic application of insights by embedding the output into your customer relationship management, enterprise resource planning, content management, or marketing automation systems.

Enterprises will never experience the full potential of data science if they continue to waste human capital on tasks that can be more effectively and efficiently performed by computers. By introducing automation and democratizing data science, companies can create entire staffs of data scientists focused on big-picture insights and solutions.

About the author

Sundeep Sanghavi, Co-founder at DataRPM Corporation Sundeep Sanghavi is a highly accomplished data junkie, innovator, and entrepreneur with more than 20 years of experience in using data as the currency to perform advanced analytics. Known for his “what if?” mindset, he co-founded DataRPM with the goal of providing a platform that delivers hyper-fast data products to organizations challenged by the volume, velocity, and variety of their big data and machine learning. Throughout his career, Sundeep has learned to productize data within day-to-day business workflows, which has led to multibillion-dollar savings. Prior to DataRPM, Sundeep founded Razorsight, a leading provider of cloud-based analytics solutions for communications service providers. Under his leadership, Razorsight experienced 11 consecutive years of growth and raised more than $30 million in venture capital before being acquired by Synchronos. Prior to Razorsight, Sundeep served in management roles at industry-leading companies, including Cable & Wireless Communications and Arthur Andersen.

  • Arun Jagannathan

    @Sundeep, I have developed few algorithms that provide a ‘paid’ search for Talent (catch n match) using a combination of Google search APIs, BigData and Data visualisation…graph search, colinearity, cohort, UX etc. With this 400k HR companies, equal number of other IP search, matrimony search and most other “search n match” enterprise can Find the next best fit easil. (Using ML they can even benefit from what key words similar-others are using for search)…need a partner/VC to gtm

You might also like...

Data Science Use Cases

Read More →