DXC Labs: Using data stories to accelerate machine learning solutions

DXC Labs has been leading the R&D for our industrialized AI offering by rapidly developing prototypes of machine learning solutions for various “data stories.” These prototypes (or minimum viable products) can be developed quickly to show customers actionable insights and engage them in deeper discussions. The prototypes, which we also call accelerators, lay the groundwork for what would become a full-blown machine learning enterprise solution.

Key to this rapid development is having a good data story.

What is a data story?

A data story describes a business situation in terms everyone (not just data scientists) can understand. In our case, the data stories involve machine-learning-based insights. One example of a data story we worked with is: “As an IT services manager, I want to proactively identify potential customer service issues so that I can intervene and optimize staff capacity/cost.”

From the story, we cull the business requirements. We then build a hypothesis designed for the business situation, get relevant data from the business stakeholders or from publicly available data, and convert the hypothesis into actionable insights that address the problem.

Why a data story is important

The data story is something that customers understand. It is a business problem or question. By using data stories, we understand the business problem and therefore can rapidly prototype solutions. Very quickly, customers “see” actionable insights from raw data.

By using data stories, we developers understand:

  1. Why we should use one algorithm over another. Because we have the end-goal in mind from the data story, we can create relevant hypotheses and appropriate algorithms.
  2. How to effectively train the algorithm. We can develop an algorithm, train it, and retrain it effectively because we know what the business issue is and what the end result should look like.

Actionable insights

For customers, our prototypes provide visualizations of the predicted results as “actionable insights” in Microsoft’s Power BI that can be drilled into for further analysis. These swift insights are powered by Microsoft Azure’s Cortana Intelligence Suite.

For developers, all common components are available in GitHub. The algorithms — for example, predicting customer churn or customer segmentation — can be reused because the algorithms have data features common across different industries.

My favorite data story

DXC Labs has built over 20 different machine learning solution accelerators for data stories that cover a range of industries.

One of my favorites is a data story on community healthcare. The data story was: “Reduce patient care cost and improve patient care and logistics for elective care.” (See figure.)

Predicting diabetes risk to reduce patient care costs — visualizations of the actionable insights obtained from otherwise disparate data. We used a machine learning algorithm for predictive analytics to deduce patterns from historical data. Click image to expand.

We segmented the population of diabetics in selected communities in order to predict the diabetes risk in new patients and schedule appointments for the most critical care patients, thereby increasing efficiencies and decreasing community care cost.  Machine learning was used to bring immediate results to the community. That was very rewarding.

In working with data stories, I am looking at machine learning with a whole new (business) perspective! And, of course, I am learning along the way.


Sukanya Kuppuswamy focuses on machine learning and advanced analytics in DXC Labs in India. A data enthusiast, her experience includes consulting and developing technology solutions for multiple business situations in the insurance industry. @sukanyaswamy

 

 

RELATED LINKS

Without a good ‘story,’ data is just a bunch of numbers

Improving Healthcare with Industrial Machine Learning

Hesitant to adopt machine learning in 2017? This may change your mind.

 

Comments

  1. Rajesh Hari Krishnan says:

    Very interesting to know that DXC labs is so active in the MI space. Will look at opportunities of engagement in the future 🙂

    Like

Trackbacks

  1. […] DXC Labs: Using data stories to accelerate machine learning solutions […]

    Like

  2. […] DXC Labs: Using data stories to accelerate machine learning solutions […]

    Like

  3. […] DXC Labs: Using data stories to accelerate machine learning solutions […]

    Like

  4. […] DXC Labs: Using data stories to accelerate machine learning solutions […]

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: