Big Data Applications

Advani

Jai Advani

Accenture

Leveraging M2M Data to Drive Operational Efficiency, Business Effectiveness, and Customer Experience

Organizations are inundated with data (machine to machine or M2M) generated from network devices, household devices and sensors enabled in consumer devices. While there had been a lot of focus on acquiring and managing this data in the last few years, organizations are now focused on the insights and intelligence that can be generated using the M2M data to improve business outcomes. This session will focus on real-life use cases that focus on developing intelligence about networks and customers using M2M data. These use cases will highlight how organizations are able to leverage M2M data and generate insights to improve operational efficiency, targeting effectiveness and enhance customer experience.

Bio

Dr. Jai Advani is a Senior Manager in Accenture Digital practice and North America Lead of Advanced Analytics practice for Communication, Media and Technology (CMT) vertical. He is responsible for expert sales, solutioning and delivery of analytics engagements for industry and functional analytics solutions. He joined Accenture in 2011 and has been instrumental in building the industry analytics solutions focused on M2M Analytics with several successful engagements across multiple clients.
Jai has more than 12 years of experience with global conglomerates focused on research and analytics. He has been a guest faculty in several management schools and leading industry conferences. Jai earned his Bachelor’s Degree in Mechanical Engineering from M.S. University, Vadodara (India) and Doctorate in Management from the Indian Institute of Management (IIM) Bangalore.

Back to schedule

Alice Albrecht

Yahoo

Best Practices in Online Experimentation

Online controlled experimentation (A/B testing) is widely used at internet companies such as Yahoo, Facebook, Google, etc. It has become the gold standard to test, validate, and evaluate product changes ranging from design to algorithm. It also produces rich data sets used in post-hoc analyses for a deeper understanding of user preference and behavior by using various statistical testing and data mining methods. While A/B testing sounds straightforward, it can be challenging in practice when applied to massive online data. The target audience for this presentation is business stakeholders and data scientists who are interested in employing randomized controlled experiments for rigorous analyses leading to data-driven decisions.

This presentation aims to provide specific guidance on best practices in online experiment design and analysis. The topics include defining experiment goals, deciding time length of test, specifying user engagement metrics, establishing pretest baseline, calculating sample size, detecting outliers, performing statistical tests, as well as conducting quality assurance procedures.

Examples from Yahoo’s practice will be used in this presentation to illustrate how we implement these best practices. The audience should leave with a better understanding of A/B testing in order to start applying the same principles to address their business needs.

Bio

Alice Albrecht is a data scientist at Yahoo! working primarily with mobile applications. She uses experimentation to help deliver data-driven insights into how to improve user engagement and retention. In addition to A/B testing, she also performs post-hoc deep-dive analyses to gain greater understanding about the particular metrics that are most important to each application. She uses her extensive background in experimental psychology and vision science to help design more effective experiments company-wide and provides consultation on possible design changes from a cognitive science perspective. Alice received her PhD in cognitive neuroscience, with an emphasis on vision science, from Yale University in 2013.

Back to schedule

Chakraborty

Goutam Chakraborty

Oklahoma State University

How Does Industry Define Analytics?  Results from a Text Mining Study of Analytics Job Postings

Since 2010, there has been a surge in academic Analytics degree programs. Academic institutions work to understand analytics needs of industry to design programs to graduate analytics professionals qualified to meet industry demands. Similarly, industries work to understand and integrate analytics methodologies from this emerging field into their business strategies. Thus, it is difficult to determine who is defining the job of the analytics professional, academe or industry.

The focus of this talk is a text mining study of analytics job postings aimed to determine how industry defines an analytics professional. Analytics job postings from relevant websites formed the dataset for the study. Standard text mining techniques were used to determine job qualifications common among analytics positions in industry including operations research-oriented methodologies and general descriptive, predictive and prescriptive modeling techniques. Distinct sets of job qualifications emerged to reveal industry’s definition of a specific set of positions for analytics professionals. (Authors: Goutam Chakraborty, Hamed Zolbanin, Melissa R. Bowers, Jeff Camm)

Bio

Dr. Goutam Chakraborty is the director of Graduate Certificate in Business Data Mining and Ralph A. and Peggy A. Brenneman professor of marketing at Oklahoma State University. Goutam is an internationally known expert in the field of data mining and analytics and has presented numerous workshops to executives and educators all over the world. He has won many teaching awards including SAS® Distinguished Professor Award and Regents Distinguished Teaching Award at OSU.

Goutam’s research has been published in many scholarly journals such as Journal of Interactive Marketing, Journal of Advertising Research, Journal of Advertising and Journal of Business Research. He coauthored two books: Text Mining and Analysis: Practical Methods, Examples, and Case Studies Using SAS® and Contemporary Database Marketing: Concepts and Applications.

Back to schedule

Dowlaty

Zubin Dowlaty

Mu Sigma

Internet of Things: Digitizing the Enterprise

Over the years, executives have been daunted by the problem of data either being unavailable or insufficient. In most cases, we have given in to this constraint and made decisions with the limited information available at hand. What if you can create the data you needed to make informed decisions you always wanted to? Internet of Things (IoT) can make this a reality by digitizing the enterprise. Breaking the constraint of data unavailability will be a game changer for enterprises in making informed decisions. This will provide the next generation competitive edge for enterprises by helping them benefit from changes in the market rather than just reacting or adjusting to them. In this session, Mu Sigma will review strategies that enterprise businesses can use to realize and sustain IoT. Also, will demonstrate interactive real-world cases that showcase how IoT can be a game changer.

Bio

Zubin Dowlaty deploys an arsenal of innovative analytics technology and quantitative techniques into the Mu Sigma ecosystem. Prior to joining Mu Sigma, Dowlaty served for eight years as VP Emerging Technology, VP Decision Sciences, and Director Consumer Insight for the largest global hotelier, InterContinental Hotels Group (NYSE: IHG). He has over 20 years of direct experience in applying quantitative methods to extract value from corporate data assets. Past responsibilities included creating and driving the global management of the data warehouse, applied business intelligence, analytics/advanced analytics, and market research departments at IHG. During his tenure the department consistently ranked in the top five of 34 global departments within IHG based on internal employee satisfaction scores. Dowlaty earned a BSBA in Finance from the University of Florida and an MA in Economics specializing in Econometrics from the University of South Carolina, and worked on a PhD MIS at University of Georgia.

Back to schedule

Gopi Vikranth

Mu Sigma

Internet of Things: Digitizing the Enterprise

Over the years, executives have been daunted by the problem of data either being unavailable or insufficient. In most cases, we have given in to this constraint and made decisions with the limited information available at hand. What if you can create the data you needed to make informed decisions you always wanted to? Internet of Things (IoT) can make this a reality by digitizing the enterprise. Breaking the constraint of data unavailability will be a game changer for enterprises in making informed decisions. This will provide the next generation competitive edge for enterprises by helping them benefit from changes in the market rather than just reacting or adjusting to them.  In this session, Mu Sigma will review strategies that enterprise businesses can use to realize and sustain IoT. Also, will demonstrate interactive real-world cases that showcase how IoT can be a game changer.

Bio

Gopi has 10 years of experience in the advanced decision sciences space with extensive experience in retail, hospitality, pharmaceutical, insurance and online advertising industries. At Mu Sigma, he works closely with Fortune 500 companies around the globe counseling them on how to institutionalize data-driven decision-making. More specifically, he is focused on accelerated scaling of analytical centers of excellence across horizontals, helping corporations leverage and monetize big data and decision sciences for top line growth and the creation of innovative methodologies using cross industry learnings. Gopi received his Bachelor of Technology and Masters of Technology in 2006 from from the Indian Institute of Technology, Bombay.

Back to schedule

Wasserkrug

Segev Wasserkrug

IBM

Wearing our Analytics: Applying Business Analytics to Wearable Technology

Wearable technology such as fitness trackers and smart watches, has the potential to transform our lives. This technology, and the resulting new data availability, can bring significant value to businesses that seek to transform their processes. However, realizing the potential of wearable technology poses significant challenges: Developing and employing suitable analytics algorithms; implementing an “analytics architecture” to correctly distribute the analytics across the multiple devices; and maintaining individual privacy.

  • Using scenarios, data, and algorithms from our work in this field, this talk will include:
  • Examples of how wearable data can enhance business processes
  • The various challenges in applying analytics algorithms to wearable data.
  • A description of the various analytical stages required to provide solutions in this domain.
  • The requirements from a wearable analytics architecture.
  • Possible solution approaches for the various challenges.

We believe that participants in this session will become as enthusiastic about this domain as we are. (Segev Wasserkrug, Neil Bartlett, Lior Limonad, Sergey Zeltyn, Alexander Zadorojniy, Nir Mashkif)

Bio

Segev Wasserkrug is a Business Optimization Leader and a Senior Technical Staff Member at the IBM Haifa Research Lab. Segev has over fourteen years practical experience in leading, developing and applying advanced optimization and analytical techniques to customer problems in a variety of domains, including workforce, logistics, scheduling, and water and wastewater operations. Segev’s current main focus is on applying analytics and optimization to mobile solutions and applications, including applying analytics to business applications of wearable technology

Segev has a strong background in a variety of areas including optimization, simulation, stochastic modeling, machine learning, and computer science. Segev received his Ph.D. in information systems engineering, and his M.Sc. and B.A.in computer science from the Technion – Israel Institute of Technology. Segev has numerous academic publications and patents.

Back to schedule