Day 1 - Wednesday, March 29, 2017

7:15 AM

Breakfast and Networking

Location: Ballroom Reception

8:00 AM

Introduction and Logistics

Location: Ballroom

8:15 AM

Keynote Speaker - Analytics Landscape

Demystifying Modern Data

Location: Ballroom
Speakers: Vikram Joshi & Rebecca Ye
It has been over a decade since the birth of Hadoop. With over $20 billion dollars in venture capital poured into open source technologies for big data, the expectation is that we are well past the hype cycle and have a mature set of products that meet the growing needs of businesses. However, reality speaks otherwise. Demystifying modern data continues to be harder than ever. And it requires programmers, a far cry from the promise of data democratization and self-service analytics. Should we have hope? What does the future hold?

9:30 AM

Break Out Sessions I

How the "Cloud" changes the Artificial Intelligence, Machine Learning & Deep Learning Conversation

Location: Salon 1
Speaker: Harsha Konduri
The 'Cloud' is finally making Machine Learning and Artificial Intelligence practical. The session explores the fundamental tenants of why Cloud Computing changes the way enterprises should theorize, architect and develop Artificial Intelligence and Machine Learning solutions. The advent of GPUs for Deep Learning, the "Uber" & "Airbnb" model of renting compute cycles and sophisticated Cognitive APIs developed by 'hyper-scale' Cloud companies that have massive ($1B+) consumer internet properties give the cloud an unfair advantage for ML & AI. Finally, the session will arm students, technologists and business leaders with a new lens of looking at Advanced Analytics in the new 'Cloud' world order.


Blueprints for an AWS Analytic Platform

Location: Salon 2
Speaker: Joel Lieser
* Immutability Changes Everything * Collect and store as much of the data generated by your organization as possible * Ubiquitous availability of this data to a multitude of applications and processes * Manage Compute and Storage Independently * Stateless, ephemeral, and horizontally scalable distributed compute * Decoupled, infinite storage * Choose Proven Open Source Solutions; Loosely-Couple Them * Low-risk investments that can be easily swapped * Multi-Tenant Platform * Tolerate and segregate disparate applications having diverse resource/compute demands There are a number of ways to successfully deploy a distributed analytic platform, and this session highlights one of those approaches. It minimizes costs while providing infinite and seamless scalability to organizations of any size.


Prescriptive Analytics & Recommendation Platform

Location: Salon 3
Speaker: Anil Daniel
We would like to present the Data Analytics platform architecture and use cases built for Cisco's eCommerce where 48 Billion USD worth of transactions get processed. This platform will enable prescriptive business insights and recommendations within the commerce transaction(deals and orders) workflow. This will also be used to make data driven decisions in the New Product Introductions process as Cisco works on new Offers in addition to key commerce KPIs. The platform uses Cassandra, Spark, Scoop, Solr, Kafka and Scala built using the DataStax Distribution. We would be excited to share how we are helping achieve these goals for Cisco.


Vision and Intelligence from the Industrial Edge

Location: The Den
Speaker: Kent Colling
“How do we get the data?” Invariably this is the first question asked by plant operators. According to IMS Research "85% of the billions of sensors and data access points needed to monitor machines exist today, but aren't being accessed." In the Industrial Internet of Things (IIoT) market we tout the value of the cloud and big data analytics, but often the reality is true insight occurs at the machine itself. In this session we present practical IIoT deployment examples, challenges, tradeoffs made, and machine health foresight gained through distributed data analytics. Distributed Analytics (FOG computing) puts data processing and continuous learning, machine intelligence at the edge of the network, in effect, breathing self-diagnosing life into the machine.


Payment Threat Analysis

Location: The Great Room II
Speaker: Atif Mohammad
The threats, which are real and our Banking and Payment Services are constantly facing as one of diverse high priority attack targets in terms of Computer Network Attacks. Such tactics utilized by Cyber-Bandits have arguably become the greatest ubiquitous medium of the threat to known CIA (confidentiality, integrity and availability) of transaction dealing with financial, retail, corporate and investment banking. Banking is the core for such breaches and the strategic threat to the payment systems and services is at the heart of these criminal activities of an interdependent financial system. This discussion will shed light, how to mitigate such threats and take preventive actions.

10:30 AM

AM Break & Expo

Location: Ballroom Reception

11:00 AM

Break Out Sessions II

Democratizing Data Science

Location: Salon 1
Speaker: Andrew Orso
As the value of data science continues to be realized and data scientists become harder to find, the need to empower employees to leverage cutting-edge analytical approaches on large amounts of data is critical for continued growth. This talk will explore a business-facing data-science tool that has been developed by the Red Ventures data-science team and how it addresses this challenge. This tool allows non-data-scientists to build and deploy real-time predictive and forecasting models as well as set up anomaly detection on data feeds, with little to no support from data scientists. Specific focus will be given to building organizational support, integration with current systems, and gaining adoption. Red Ventures use cases in the customer acquisition funnel, including customer churn, agent staffing, and sales call coaching, will be used for illustration.


Transformation in financial services and cloud content management

Location: Salon 2
Speaker: Tom Shepherd
Transformation in financial services and cloud content management


Enterprise Data Strategy

Location: Salon 3
Speaker: Shikha Kashyap
Breakout Session Description: When building an enterprise data strategy, consider “why?” before "how". Data is the most powerful, yet underutilized and poorly managed organizational asset. What are the best practices for developing the data roadmap. Take baby steps and quick wins as part of putting together the data roadmap. It is really easy to get lost in the technology stack evaluation before data goals are established within the organization.


The Urge to Review Algorithms: A Technical-Contextual Framework

Location: The Den
Speaker: Reza Mousavi, PhD
Algorithmic decision making is spreading into our businesses, governments, and schools. While commentators are demarcating their public relevance, algorithms are now making decisions in areas where human judgment and discretion traditionally played a big role. In the city of Chicago, predictive algorithms are being used in a wide range contexts including identifying individuals who would likely commit a violent crime to identifying restaurants that would likely violate health inspection rules to predicting rodent population. Given the widespread use of algorithms in a wide array of contexts, there is a need to ensure that these algorithms are first acting as intended. This refers to a review from a technical perspective, such as meeting performance criteria. The second is ensuring that models comport with our established norms. For example, ensuring that a model does not operate in biased or discriminatory manner. In this session, we provide an overview of our framework for reviewing algorithms and models. Our focus is the machine learning models developed by data scientists. The framework is illustrated using a case study of an algorithm designed by the City of Chicago for food inspection. The result is a review process that considers both the technical aspects of the models as well as how a model comports with our established norms. This can be used by practitioners and policymakers in developing and vetting algorithms.

12:00 PM

Lunch & Expo

Location: Ballroom Reception

1:00 PM

Break Out Sessions III

A World with Watson

Location: Salon 1
Speaker: Nicole Bradley
Augmenting Intelligence with Watson


Case Study: Strategic impact of analytics and insights at a healthcare system

Location: Salon 2
Speaker: Matthias Krebs
The presentation takes the audience on a 2 year journey in which insights & analytics transformed how decisions are being made at a large healthcare system. The new fact-based approach resulted in an award winning TV advertising campaign and a drastic ROI improvement. Both causing executives to see marketing no longer as a cost center but instead as a profit center.


Designing a Modern Data Warehouse + Data Lake

Location: Salon 3
Speaker: Melissa Coates
Join us for a discussion of strategies and architecture options for implementing a modern data warehousing environment. We will explore advantages of augmenting an existing data warehouse investment with a data lake, and ideas for organizing the data lake for optimal data retrieval. We will also look at situations when federated queries are appropriate for employing data virtualization. This is an intermediate session suitable for attendees who are familiar with data warehousing fundamentals.


Marketing Science for Customer Centricity

Location: The Den
Speaker: Michael Scully
Customer centricity is often misunderstood. Changes in customer behavior and enabling technology have given rise to increasing focus on customer centric marketing in recent years, yet transformation of marketing techniques presents a number of challenges. In this talk, Michael Scully discusses the "what, why, and how" of customer centricity and how predictive analytics for customer lifetime value (CLV) can improve quantification of return on marketing investment.

2:00 PM

Afternoon Break & Expo

Location: Ballroom Reception

2:30 PM

Break Out Sessions IV

Data Science at Scale – Applications in Digital Marketing

Location: Salon 1
Speaker: Kevin Pedde
Paid Search is a popular digital marketing channel that allows companies to reach audiences based on what they are searching for. When done correctly, paid search can significantly impact strategic initiatives such as customer acquisition and lead generation. At a high level, the task is simple – decide which keywords to bid on and how much at each time to maximize value. However, the complexity of performance by day-of-week, time-of-day, device, demographics, and geography combined with rapidly changing auction dynamics for hundreds of thousands of potential keywords makes this task daunting. This talk will focus on Red Venture’s NCTA award winning technology (http://nctechnology.tumblr.com/post/153222113412/2016-nc-tech-awards-recap) to address this problem, Sematic. Sematic builds tens of thousands of predictive models and places more than half a million distinct bids every day. Specific focus will be given to the underlying data science infrastructure required to support the tool and how analysts interact with it. Specific results from Red Ventures customer acquisition funnel will be used for illustration.


Increasing web and social traffic with education-industry partnerships

Location: Salon 2
Speaker: Robin Blom
Cardinal Metrics is a student-run media analytics agency housed in the campus newsroom of Ball State University. The organization does web and social media analytics consulting for internal and external clients. One of those clients is Visit Indy, the bureau that promotes the city of Indianapolis with its multi-million dollar budget to attract more business for hotels, restaurants, stores, and entertainment venues. The partnership between Cardinal Metrics and Visit Indy started as an immersive-learning project for a “Media Analytics in Practice” course. A dozen undergraduate students from journalism, public relations, and advertising worked in two teams to advise Visit Indy on its web and social media strategies based on analyses of all available data that the organization continuously gathers through all its channels and platforms. Visit Indy has already implemented several recommendations made by the students to attract more visitors to its website and to create larger name recognition. This presentation will demonstrate how partnerships between higher education and the industry can be highly beneficial for everyone involved—providing students with transformative learning experiences and companies with innovative avenues to increase engagement, exposure, revenue, and profit.


Unlocking Open Data's Network Effect

Location: Salon 3
Speaker: Leonard Fishman
Open data can help us rise to humanity's toughest challenges, but only if we maximize its network effect. To do that, we need both data producers and data users to work collaboratively to create metadata-rich datasets. 18M open datasets exist today, and growth is accelerating. Data producers are uniquely positioned to create foundational metadata that can make open data more accessible. This includes establishing clear provenance, providing a data dictionary, identifying relationships that resolve ambiguity and irregularities between associations within the data, and documenting granular attribute-level information. Data producers cannot anticipate every use of a dataset, especially as data frequently has uses in other sectors and industries. As data users – both humans and computers – normalize, extract meaning, and identify correlations between datasets within their industries, they are creating additional metadata. However, this work too often resides in silos. It is used for one project, then lost forever, only to be repeated from scratch by the next person to touch the data. If data producers published their data using common taxonomies and architectures and if data users had a platform on which they could discuss a dataset, publish their analysis and queries, and in general enhance metadata, data would pass through the complete metadata maturity model and data’s network effect could be unlocked.


Preparing for Data Informed Decision Making in a Small Healthcare Organization

Location: The Den
Speaker: Sarah Kruse
Big Data is often associated with big organizations, but what happens when a small organization is faced with requirements to capture, use, and share data?  There are many untold stories about the successes and challenges faced by community health centers since the 2004 mandate for healthcare organizations implement electronic health records.  This presentation tells one of those stories.  Participants will gain an understanding of the federal reporting requirements for FQHCs, learn about goals to better understand patient clinical outcomes, costs, trends, and utilization patterns, and hear about organizational successes and challenges to achieving data informed decision making related to population health and reducing health disparities.


Biomarker Discovery in Sjögren’s Syndrome: Database Analysis and Web Interface Design

Location: The Great Room II
Speakers: Nirav Shah & Mohamed Shehab, PhD
Identification of candidate biomarkers is crucial for the development of novel therapies against diseases. Access to the freely available meta-data for various diseases generated in past 15 years using micro-array technology is difficult. In this study, we present a simple web interface for the easy retrieval of gene expression profiles of queried genes from 8 curated databases associated with multiple Rheumatic diseases and various tissue sites. This web-tool allows researchers to retrieve meta-analysis results showing various parameters such as gene names, fold changes and p-values. In our talk, we will highlight major features, specific challenges related to the development of this interface and future directions for further improvement.

3:45 PM

Break Out Sessions V

Business Strategy Analytics: Integrating Reason and Data

Location: Salon 1
Speakers: Victor Chen, PhD & Mikael Hagsroem
This 60-minute breakout session is to present the audience with some state-of-the-art knowledge and practice on business strategy analytics by both a practitioner, Mr. Mikael Hagstroem (Partner and COO, McKinsey Analytics), and an academic researcher, Dr. Victor Zitian Chen (Assistant Professor of International Management, Belk College of Business, UNC Charlotte). This session will focus on how business strategy can benefit from integrating scientific research and big data to strive for managerial performance. The session will start with Dr. Chen’s introduction of the background of the session and the keynote speaker. Drawing on his role as COO of McKinsey Analytics and formerly Executive VP of SAS, Mr. Hagstroem will give a keynote speech on how to create a data-driven culture within an organization. He will focus on which companies are renowned for being data-driven and what attributes make them truly distinctive. He will also discuss the risks and how you mitigate those risks. After that, Dr. Chen will give a presentation on how academic research based on science-based principles and rigor can enhance analytics in business decisions. He will showcase two of his working projects as examples: one on performance measurement system of corporations that enhance the benefits for all stakeholders; another on entry mode choice to enhance post-entry performance in cross-border M&As. The last 20 minutes will be devoted to a genuine interaction among the speakers and the audience, focusing on how business and academic researchers can collaborate to improve each other’s work, and identifying gaps in the current research in business analytics.


4 Visions for IoT

Location: Salon 2
Speaker: Peter Darragh
Microsoft’s “Internet of Things” platform is used by many companies but most of the case-studies understandably concentrate on publicly-owned global giants and ignore how the platform can be utilized by smaller organizations. Using the experiences of four privately held organizations the session confirms that Microsoft’s IoT suite is applicable to a variety of different industries, a variety of applications and a variety of pre-existing technology assets. By reviewing the experience of those organizations the session describes some common traits and provides general advice on how smaller organizations can succeed with Microsoft’s IoT platform.


Data-Driven Marketing

Location: Salon 3
Speaker: Jake Finkelstein
This workshop will dig into how data-driven marketing improves results and enhances the accountability of the marketing department to the C-suite. The first part of the session will explore how qualitative and quantitative data can be gathered and used to inform marketing strategy and tactics; using case study examples and resources to guide practical application of the methodology. The second half will be a hands on learning session where your real-world questions will be answered by Jake.


VALSE- Visual Analytics for Large Scale Ethnography

Location: The Den
Speakers: Chris Beokrem, Donna Lanclos, & Eric Sauda
VALSE is a framework for providing access to stored long term, large scale human behavior which allows an expert to view avatars representing real-time motion and also visualizations of motion summaries, all controlled using familiar DVR-style interface. VALSE requires robust algorithms for human tracking and activity analysis from video and usable interfaces for domain experts to query and understand the data. Our system combines tracking and activity recognition methods into an interactive, real- time system for visual data analysis in the study of motion behavior. VALSE combines time line visualization with spatial visualizations linked and open to exploration by ethnographers or other domain experts. A critical part of the VALSE interface is the methods of identifying meaningful behavior as the intersection of time, people and objects; the system can then search for all incidents of this behavior over very long time periods. This approach leverages the ability of human ethnographers to recognize meaningful patterns of behavior with the computational ability to apply these insights for very extended periods of time.

5:15 PM

Expo

Location: Ballroom Reception