Tuesday, May 10, 2016

WELCOME & KEYNOTE - How Statistics (And a Little Public Data) Can Change a City

The creator of I Quant NY, a data science and policy blog that focuses on insights drawn from New York City's public data, Ben Wellington, advocates the analysis of open data to affect policy. The Open Data movement is growing, and governments are releasing vast amounts of data to the public. As citizens push for more transparency, it is fair to ask what we can actually do to derive actionable insights from this data. How can this data help us improve the cities we live and work in, whether we are policymakers, businesses, or residents? Wellington explores how he's used his blog and some simple data science techniques to make changes in New York City. He discusses best practices for data science in the policy space, explores how storytelling is an important aspect of data science, and highlights the various datadriven interactions he's had with city agencies. He contends that data science need not use complicated math: It's often more about curiosity and the questions we ask than the complexity of the equations we use.

Ben Wellington, I Quant NY

COFFEE BREAK in the Data Solutions Showcase

A102: New Approaches to Data Management

The speed at which global business operates today is fueling a demand for systems that can enable insight about everything from sentiments to machines in the moment. Real-time data systems are offering the ability to accelerate business reaction time as never before.

How to Count Tens of Billions of Daily Events in Real Time Using Open Source Technologies

Twitter’s 300 million-plus users generate tens of billions of tweet views per day. Aggregating these events in real time—in a robust enough way to incorporate into our products—presents a massive scaling challenge. This talk introduces TSAR (the TimeSeries AggregatoR), a robust, flexible, and scalable service for real-time event aggregation designed to solve this problem and a range of similar ones. The discussion covers how Twitter built TSAR using Python and Scala from the ground up, almost entirely on open source technologies (Storm, Summingbird, Kafka, Aurora, and others) and describes some of the challenges Twitter faced in scaling it to process tens of billions of events per day.

Claudia Perlich, Chief Scientist, Dstillery

From Data Management to Data Driven

With the advent of machine learning, cognitive computing, graph technology, IoT, and data marketplaces, decision management will never the the same again. Khanna explores how emerging trends such as cloud, mobile, social, and big data analytics are making data management a disruptive technology in this new age of the consumer. He reviews some of the cutting edge use cases in data management making a real impact on our lives and introduce possibilities for the future.

Ajay Khanna, Vice President, Product Marketing, Reltio

A103: Supporting Modern Applications

Organizations are increasingly seeking to use their data for business advantage. However, they often find difficulty in leveraging data as fast as they would like. New data management technologies and techniques help organizations meet the changing requirements posed by modern applications for speed, scale, and flexibility.

Eliminating the Data Constraint in Application Development

Many consider the No. 1 bottleneck in application development to be the provisioning of parity data from production data sources to use in development environments, QA, UAT, and integration testing, including building and managing development and QA environments. This session explores what the impact is, as well as how to solve this bottleneck with virtual data and define metrics to track to validate the solution.

Kyle Hailey, Technical Evangelist, Delphix

The New Stack for Content Managmenet: Supercharging the Content Repository with Elasticsearch and MongoDB

We strongly think SQL and NoSQL technologies should be used side by side to deliver a storage solution that can scale to the sky while preserving data security, and guaranteeing transactional write operations. Nuxeo is an open source platform centered around a content repository. Over the years, our use of the technologies has evolved. Among other changes, our content repository migrated from an Object Database to SQL. For a couple of years, we’ve been leveraging NoSQL technologies, enabling us to provide a full range of solutions, including a new hybrid SQL + NoSQL architecture, depending on the challenges and the requirements of each project. This presentation explains the technical and design choices we had to make to keep the content repository up-to-date and provide incredible performance to our users.

Vladimir Pasquier, Senior Software Developer, Nuxeo

A104: Going Beyond Relational

The era of relying on one relational database system is winding down. Today, data is flowing into organizations from new sources outside of and within organizations. The question is: How can companies leverage these new data types as they seek to go beyond relational?

An Introduction of Big Data Concepts for Relational Users

Big Data cannot be ignored if you wish to continue pursuing a data management career. But what is Big Data? Join this session to find out how Big Data technologies differ greatly from Oracle, SQL Server, DB2, and other relational database systems.

Craig S. Mullins, President & Principal Consultant, Mullins Consulting, Inc. IBM Gold Consultant

NETWORKING RECEPTION in the Data Solutions Showcase

B102: Becoming a Data-Driven Enterprise

Data-driven enterprises have more information than ever available for scrutiny. In today’s fast-paced digital economy, it is understood that effective data management strategies can have significant impact, and leading vendors are providing innovative products and services aimed at helping organizations derive more value from their data.

The Two Faces of Data Strategy: Why It Can Be Hard for Business and IT to See Eye to Eye

There is no doubt in the boardroom or throughout the organization that data is a valuable asset, but despite the ardent agreement about its value, business and IT leaders alike have distinct differences of opinion on how data should be managed and used. In this session, we look at the heart of the data strategy arguments, share why the different mindsets exist, and provide methods to develop an enterprise approach that allows for optimized, efficient data management and maximizes business value generation opportunities. It is important to understand what elements of data strategy are best under the innovative wing of the business and which portions need the careful administration and oversight of IT.

Anne Buff, Business Solutions Manager, SAS Best Practices, SAS Institute

Move the Data That Moves Your Business

To manage growing volumes, varieties and velocities of data today, and business demands to use it, you need to rapidly enable data for analytics. But integration across data warehouse, Hadoop, and other platforms, on premises or in the cloud, can be prohibitively complex and costly. Traditional data movement and transformation tools are slow and require manually coded commands that tie up your best programmers. Learn how you can ingest data at high speed into modern data warehouses with automation that accelerates analytics projects, enables timely insights, and improves agility.

Kevin Petrie, Senior Director of Marketing, Attunity, Inc.

B103: Big Data Analytics in Action

There are a myriad of ways Big Data is being used to measure things that were previously unmeasurable, anticipate events that could never before be foreseen, and analyze data that never before could have been collected. Today, Big Data analytics is having real-world impact.

A Data-Driven Approach to Environmental Regulation at U.S. Environmental Protection Agency (EPA)

It was only 45 years ago that U.S. cities had smog-filled skies, and rivers were so contaminated, they burst into flames. Today, the EPA collects toxic chemical data released by industries and monitors our air/water/land using sensors for ensuring healthy and sustainable environment for our citizens. In this session, Robin A. Thottungal, the EPA’s first chief data scientist, highlights the how the agency is becoming more data-driven and shares some of challenges and the innovative solutions taken by the agency in implementing real-time monitoring of environmental parameters for gaining better understanding of the current state of our ecosystem.

Robin Augustine Thottungal, Chief Data Scientist, U.S. Environmental Protection Agency

ThoughtSpot's Presentation To Be Announced

Mike Booth, Analytic Search for Big Data, ThoughtSpot

B104: Supercharging Your Marketing With Big Data

There are more ways than ever to reach customers with marketing messages and also for them to talk back to you about what they like and what they don’t. Big Data technologies are increasingly being put to big use in marketing today to collect, analyze, and measure everything from sentiments to webpage clicks.

Using Big Data Analytics to Optimize Email Marketing Campaigns

This session takes you behind the curtain to unveil brand new (and incredibly comprehensive) email engagement insights uncovered by Constant Contact's data scientists. Compiled from the analysis of the billions of emails Constant Contact customers send every year, these insights dig deep into the data behind what factors influence subscribers to sign up, open, click, and convert. The session proves invaluable for data scientists who are looking to expand their teams’ value into areas of marketing, data vendors looking for fresh ideas on how their services can be innovatively applied to other areas of their clients’ companies, and, of course, marketers looking for the latest best practices for their own email campaigns.

Matt Laudato, Director, Big Data & Analytics, Constant Contact

NETWORKING RECEPTION in the Data Solutions Showcase

C102: Spark for Oracle Developers

This session covers moving from an Oraacle-only pipeline to a real time Spark pipeline.

Anant Asthans, Big Data Consultant/Data Scientist, Data Scientist, Pythian

C103: Integrate Big Data with Master Data on Hadoop and Spark

Big Data analytics furnishes actionable information to help business decisions—e.g., "Which of our products got four stars rating on social media in the last quarter?” Producing actionable information requires both Big Data i.e., facts, usually stored in HDFS or NoSQL DBMS as well as Master Data i.e., products, customers, usually stored in Oracle database. You either ship Master Data to Hadoop (ETL) or access it directly at the storage location through a storage handler. This technical session demonstrates Oracle Table Access for Hadoop and Spark (OTA4H) which turns Oracle database tables into Hadoop or Spark datasources thereby allowing Hive SQL or Spark SQL queries and joins without shipping data around.

Kuassi Mensah, Director of Product Management, Oracle Corporation

C104: Hadoop Data Store and Retrieval: Perceptions and Reality

Hadoop-based data stores have attracted the imagination of technologists, management, and technology leaders. Hadoop data stores are distinctly different from traditional structured relational databases. They have opened the door for new paradigms in storing structured and unstructured data and led to a concept of "store first, think later" as to how to use it. This session focuses mainly on how data can co-exist between traditional RDBMS and emerging data types within the Hadoop ecosystem. It discusses how data formats and query engines differ from the traditional access patterns. Many complex data management issues within Hadoop exist. With many open source projects running parallel, there are new ideas, new query modules, and new storage methods evolving. We discuss perceptions and realities from a data administrator and data manager perspective. 

Madhu Tumma, Director, IT Engineering, TIAA-CREF

NETWORKING RECEPTION in the Data Solutions Showcase

H102: Querying Hadoop and NoSQL Data Stores

The pressure is growing for organizations to react faster to changing opportunities and risks by using data to improve decision making. Hadoop and NoSQL data stores provide new options for organizing and aggregating data in all its forms. Find out what you need to know about querying Hadoop and NoSQL data stores.

De-Siloing Data Using Apache Drill

Study after study shows that data scientists spend 50–90% of their time gathering and preparing data. In many large organizations, this problem is exacerbated by data being stored on a variety of systems, with different structures and architectures. Apache Drill is a relatively new tool that can help solve this difficult problem by allowing analysts and data scientists to query disparate datasets in-place using standard ANSI SQL without having to define complex schemata or rebuild their entire data infrastructure. This session introduces the audience to Apache Drill and presents a case study of how Drill can be used to query a variety of data sources.

Jair Aguirre, Lead Data Scientist, Booz Allen Hamilton

H103: Harnessing Big Data With Spark

Apache Spark, an engine for large-scale data processing, can be complementary to Hadoop but it can also be deployed without it. Find out what Spark offers and why it is gaining ground in the world of Big Data.

Apache Spark and Effective Machine Learning

The introduction of Hadoop MapReduce (MR) allowed the application of algorithms to data of unprecedented scale using systems built from cheap commodity hardware. However, MR is slow, significantly curtailing its applicability to advanced iterative machine learning (ML) algorithms. These algorithms frequently need to be run multiple times in order to effectively train and optimally parameterize. Spark changed this, and, by providing speedups of 100X or more, fundamentally introduced the possibility of applying ML to Big Data and extracting meaningful insights in actionable timeframes. This presentation provides an overview of the framework Alpine Data developed and present results for a variety of well-known datasets, illustrating how the software can significantly eliminate the repetitive, trial-and-error nature of today's data science and reduce the time to an effective model.

Lawrence Spracklen, VP, Engineering, Alpine Data

H104: Building Hadoop Applications

Building Hadoop Applications

Whether Hadoop becomes the de facto management platform of the future for Big Data or simply a key component in a hybrid architecture comprised of numerous technologies, by now it is certain that Hadoop, along with its larger ecosystem, is no fly-by-night technology. Find out the key issues involved in leveraging Hadoop for Big Data applications.

Building Scale-able Machine Learning Applications on Apache Spark

There is a growing demand in the industry for highly scale-able data processing platforms that are built around simplicity of architecture, compatibility and robustness. Apache Spark is one such example, and is one of the most exciting emerging technologies in today's computing landscape. It is finding many exciting use cases, one of them being Machine Learning. In this session, a brief description of the Apache Spark Architecture will be presented, along with a case study of a machine learning algorithm to demonstrate the versatility of the Spark Architecture.

Abhik Roy, Engineering Principal, Wells Fargo

NETWORKING RECEPTION in the Data Solutions Showcase