What Is Your Data Science Super Power?

This post is authored by Wee Hyong Tok, Senior Data Scientist Manager, and Danielle Dean, Senior Data Scientist Lead, at Microsoft. Wee and Danielle are speakers at the upcoming Microsoft Data Science Summit on September 26-27 in Atlanta, GA.

How do businesses and data scientists turn raw data into intelligent action?

Why do some companies drown in volumes of data, while other companies thrive on turning data into golden strategic advantages?

In our customer engagements, interactions with data scientists, and conversations with the data science community, we repeatedly hear these questions and more:

  • “What is the secret sauce that gets you successful outcomes in your data science projects?”
  • “What tips and tricks do you use when tackling challenging data sets?”
  • “Are there techniques that will help me establish a good initial baseline model, that I can then refine?”
  • “I seem to be doing lots of trial and error when working on my model, in order to get good initial results. How can I do this more efficiently?”

Each day, data scientists tackle some of the most challenging problems in their respective industries. And in each of these data science projects, existing methods such as data exploration, feature engineering, training, evaluating, and operationalizing the model evolve to meet the requirements of the project, and new techniques are invented that advance the state of the art in data science.

When we capture the keys to success at these projects, and leverage them in our next project, we call this finding your “data science super powers.”

In the “What is Your Data Science Super Power?” talk at the Microsoft Data Science Summit we will share the stories behind several of our data science projects, including our work on:

  • Sloan Digital Sky Survey data, to understand the galaxies that make up our universe.
  • Decoding signals from the brain using electrocorticography (EcoG) signals.
  • Data from airplane engines, to improve the efficiency of manufacturers’ aircraft.
  • Customer churn data in the telecommunication industry.

Although these projects differ in many ways, the goal of each is clear: To figure out how to use data and transform it into intelligent action.

We have been working with some of the best data scientists in the community to learn and co-create techniques to do data science simpler, faster, and more efficiently. From these conversations and practical experiences, we have crystallized five data science “super powers” that you can put into action in your next project.


Join us at the Microsoft Data Science Summit on September 26-27 as we share these data science super powers and show you how to distill raw data into big ideas and big results! We plan to help you uncover your own data science super powers! 

Wee & Danielle