Data Science and Marathon Running Both Require Training and Patience

August 30, 2018
Uncategorized

Guest post by analytics and IT expert, Sundaresh Ramanathan. Not only am I a friend of Sundaresh, but I moonlight as a lover of data solutions myself. Plus, I completed a full marathon in 2018 and I think his analogy is spot on. So, I had to showcase this post.

If you got value from this 4-minute read please listen to his 37-minute  podcast interview here with Outcome Studio. -Aaron

Data Science is a Marathon – Pace Yourself

I recently, started running with the idea of running a marathon some day. Not an immediate goal, but maybe some time in the near future. While looking for running tips, I came across this great article on running a marathon. I immediately saw the similarities with any data science project.

The benefits of data science in the enterprise are undisputed. Gartner estimates that 90% of large companies will have a chief data officer to build a company-wide strategy for managing, leveraging, and securing data being produced worldwide each day.

Everywhere you turn, both business and IT talk about data science. But there’s also uncertainty about how to get started, especially in the context of achieving an organization’s business goals and objectives beyond the limited departmental experimentation.

Most runners will tell you, running a marathon is an undertaking that should not be taken lightly. No matter what your motivation, preparing for this event requires vision, planning and training.

Introducing data science in your organization should be like a marathon. You don’t just run a set of queries against your data and expect to see miraculous results. You need to have a vision, goal, objectives, and plan the execution to insure reliable and consistent results. NIH recently published their strategy for introducing data science, and their first step was to document their definition of the term. Here’s how NIH defines data science:

 

“The interdisciplinary field of inquiry in which quantitative and analytical approaches, processes, and systems are developed and used to extract knowledge and insights from increasingly large and/or complex sets of data.”

 

In this definition, they set their vision, goal and objectives. Here are my thoughts on how you can train for a Data Science Marathon.

Find a Compelling Reason as Fear Will Replace Excitement

If you decide to run a marathon because you bought a pair of high-tech running shoes or the new smart watch, you’re more likely to fail. Find a larger than life reason to run the marathon and you will find the motivation to keep at it.
Data Science should be about achieving larger than life business results with minimal effort, not about using the latest technology. This may seem obvious, teams often embark on an analytics project by picking the latest technology or tool available. Data science is about storytelling. Focus on telling stories with numbers. The art of storytelling is simple and complex at the same time. Stories provoke thought and bring out insights often overlooked.

Foster User Engagement, Plus Not everyone Will Be Supportive

When you make a big decision in your life, you want your friends and family in your corner. Of course, not everyone in your life is always going to be supportive. Find the motivation from within to keep going when others are suggesting you should stop.

Data science involves everything related to data preparation, cleansing, and analysis. Thus, Data Science combines data experts, system experts, business experts, leaders and users. It combines their knowledge and passion for storytelling to arrive at ingenious ways to narrate a story. It can be a very complex process and not everyone will be supportive.

Health Issues Are Common and You’ll Uncover Dirty Secrets About Yourself

Twenty six miles is a long way to go. It’s completely normal to have shoes with plenty of cushioning that are great for your needs, yet still they can feel terrible after your long runs.

Naturally, data is the raw material of data science. Generally, the more data you have and the more varied data sources are, the better. So go beyond traditional systems, and be on the lookout for interesting and potentially relevant new data sources, such as system logs, customer data, business metrics, social media data, sensor data and/or any data that is relevant for impactful storytelling.

Note though, all this data may reveal some not so flattering facts about your business or product. Don’t let this stop you. This is not unique to your project, most transitions to analytics driven decision making will lead to more opportunities for improvement than planned.

Sh*t Happens: It’s a High-Risk Project

When you first start running long distances, your body will experience a lot of firsts and it may take time for it to adjust to everything that is happening. You may find that foods you formerly tolerated with ease don’t agree with you when you have a long run the next day.

Data science is one part experimentation. You are continuously trying new things. Looking to answer questions no one has answered before. You will inevitably hit a point where you find out what you hoped for just doesn’t seem possible. Either data doesn’t actually exist, it isn’t stored in a useful format, or it’s not stored in a place you can access. Data Science is typically hyped up to be full of easy wins, but in reality that isn’t always the case.

It’s Expensive

From running shoes to healthy foods, running a marathon is not cheap. The costs add up fast.

Similarly, data science projects can be expensive and needs commitment from several facets of the organization. This costs time and money. A typical project may need commitments from data scientists, business analysts, data analysts, statisticians, and software engineers. This, in addition to the investment needed in data storage, compute, and infrastructure horsepower.

Everything Can Go Wrong During 26.2 Miles

A data science project is not different. You have a great story, vision, plans, resources and commitments. You have the best team and picked a not too large, not too small project, yet not everything may go according to plan. You didn’t meet all the objectives you set out to accomplish. This may feel like a failure, but it is not. Any time you can transition to data driven decision making, no matter how big or small, is a victory. Learn your lessons, dust off yourself, and start your next project.

 
 

Find Sundaresh online: