Annika Jimenez on Disruptive Data Science at the Strata Conference

October 29, 2013 Paul M. Davis

annikaDuring a presentation this morning at the Strata Conference in New York City, Pivotal’s Annika Jimenez outlined Pivotal’s vision of data science as a disruptive force and discussed requirements for data science success. With Visa’s Ravi Devireddy, Jimenez presented a real-world case study on Pivotal’s work with the company to apply data science to its cyber security efforts.

To set the stage, Jimenez began with an explanation of Pivotal’s definition and vision of data science. She defined the process as “The use of statistical and machine learning techniques on big multi-structured data — in a distributed computing environment — to identify correlations and causal relationships, classify and predict events, identify patterns and anomalies, and infer probabilities, interest, and sentiment.” The key elements of this definition regard Big Data with varying degrees of “structure” being leveraged within a distributed computing environment for sophisticated analytic intent.

In Jimenez’s view, the goal of data science is not simply to create analytics dashboards, or even “cool visualizations, custom querying, decision enablement, or insights.” Though these are important elements, the end goal of data science — and where true ROI lies — is found in “driving automated, low latency actions in response to events of interest.”

Screen Shot 2013-10-29 at 1.22.41 PM

An example of this in action can be found in the work Pivotal did showing that customer care call logs were useful in predicting potential to churn. In this case, the goal was to “enable systematic response to user-level likelihood to churn. If your call topic triggers an increase in likelihood to churn, let’s send an email, let’s offer a discount, let’s remove some charges —whatever, but let’s try to lower the likelihood to churn.”

Herein lies what Jimenez described as “the essence of data science — predictive modeling and supervised & unsupervised learning techniques using big data tools on a distributed computing stack.” From this process emerge models that respond to events of interest, such as those likely to trigger customer churn, that can be operationalized into automated actions.

Jimenez acknowledged that operationalizing such a process is a major undertaking and requires a significant shift in thinking and operations. “Doing this is not a cake-walk and it taxes a company’s legacy & status quo,” she said. “It’s disruptive. When companies were built to do something else entirely, getting to this requires a bit of a transformation.”

Using Yahoo’s personalized advertisement and content serving as an example of operationalized data science in action in a company without such legacy restrictions, Jimenez emphasized that these innovations are quickly transferring onto traditional enterprise operations. “Data will be the basis of competition,” she said. “If you don’t curate, enable data science, model operationalization, and app integration well…you will lose ground to the companies who are.”

Screen Shot 2013-10-29 at 1.13.48 PM

Helping companies respond to this paradigmatic shift and operationalize their data science efforts is the core of Pivotal’s mission, Jimenez explained. With a portfolio of products spanning apps, data, and analytics, Pivotal aims to help enterprises modernize and accelerate each step of a valuable and increasingly essential business cycle:
1) Apps power businesses, and those apps generate data
2) Analytic insights from that data drive new app functionality, which in-turn drives new data
3) The faster you can move around that cycle, the faster you learn, innovate & pull away from the competition

Visa’s Ravi Devireddy joined Jimenez to provide a real-world example of this cycle in action, speaking of the challenges, opportunities, and ROI involved in the company’s engagement with Pivotal to operationalize data science as part of Visa’s cyber-security efforts.

To shift towards such capabilities, Jimenez explained, companies across sectors need to address legacy concerns in apps, data, and analytics. Pivotal’s “dream team” of data scientists and engineers boast domain-specific knowledge and skills that allows the company “to support modeling efforts across any sector,” she said, including energy, retail, financial services, life sciences and healthcare, manufacturing, and communications.

Pivotal continues its presence at the Strata Conference in New York City tomorrow, Wednesday October 30th, with a Plenary Keynote at 9:05am by Josh Klahr on Designing a Data-Centric Organization.

About the Author

Biography

Previous
From Measuring to Understanding
From Measuring to Understanding

Starr Horne, a co-founder of Honeybadger.io, discusses some easy wins for collecting metrics and gives an o...

Next
Improved Java Tooling for Cloud Foundry
Improved Java Tooling for Cloud Foundry

Pivotal has released several new components that enable developers using Java, Groovy, and other JVM langua...

×

Subscribe to our Newsletter

!
Thank you!
Error - something went wrong!