Joint Presentation: Advanced Analytics at The Dow Chemical Company
Presented by: Ameya Dhaygude, Advanced Analytics, Data Scientist, The Dow Chemical Company Tim Licquia, Advanced Analytics, Data Scientist, The Dow Chemical CompanyThe Dow Chemical Company is a world leading enterprise in the chemical manufacturing industry that specializes in delivering chemical, material, and agriculture solutions to customers worldwide. Dow’s expansive global footprint encompasses 56,000 employees --- spanning 189 sites across 34 countries --- generating over 7,000 product lines for its customers, which themselves span 175 countries and nearly a dozen different commercial markets. Dow consistently ranks as a Fortune 100 company, with its revenue predominantly driven by business-to-business sales, and only a few markets where its products are sold directly to end users.This session will focus on the interesting opportunities and business challenges that impact a company of such scope and magnitude. Data scientists Tim Licquia and Ameya Dhaygude will illustrate the current analytics landscape within Dow by highlighting the various aspects of Dow’s complex operations model that leverage data science, as well as the diverse community of data scientists that support this. This will conclude with a case study on hierarchical time series forecasting, which has proven invaluable for generating actionable insight within Dow’s extensive and complicated business structure.
Defining Strategies to Integrate Analytics with Business Intelligence Platforms
The integration of Analytics and Business Intelligence platforms can be a challenging task for many organizations. Business Intelligence platforms (e.g., Data Warehouses) are often viewed as old solutions in need of an overhaul. Conversely, Analytics platforms (e.g., Data Lakes) are perceived as their replacement.In this session, participants will learn strategies to integrate, not replace, legacy Business Intelligence with Analytics platforms. One primary goal of this session is to show that Data Lakes and Data Warehouses can be complementary, not competing solutions.
Embedding Predictive Analytics into business processes to drive operational value
While Predictive Analytics adoption has grown steadily for several years, the overall CAGR has fallen short of lofty estimates. Most organizations acknowledge the need to leverage Predictive Analytics for improved decision-making, and businesses have made significant investments; so why has the industry not exploded? Join us to learn how QueBIT has delivered massive ROI for our clients through a unique approach to implementing Advanced Analytics solutions in areas such as Demand Planning and Price Optimization. This approach ensures organizational goals are achieved and adoption is strong.
Networking Coffee Break
Understanding Customer Behavior with Predictive Analytics & a 360 Degree View of your Customers
Knowing what motivates your customers unlocks an abundance of information and the customer 360 is the central point of understanding your customer. In this session, Rich will share how a Customer Centric View approach was utilized in designing the Customer 360 to grow sales and improve margins along with the specific initiatives of the Customer 360. The session will cover the predictive models utilized to understand the drivers and root cause of sales performance, how to better target customers for acquisition, better understand customer churn and proactively manage retention.
Avoiding Bias in Predictive Analytics
For those of us who have the ever exciting and growing task of working with Predictive Data models to help solve some of organization’s biggest inefficiencies, questions, or problems, perpetuating bias is a way too easy-to-make mistake, and we should all be familiarized with it by now. During this session we'll look at some examples of bias in predictive models, and analyze different ways in which we can prevent and correct it.
Deep Learning Fast Start
Deep Learning has emerged as one of the most promising areas of AI and Machine Learning. The presentation will provide a "fast start" introduction to deep learning and artificial neural networks. We will begin with a conceptual understanding of artificial neural networks, provide a gentle introduction to the main technical concepts, and conclude by creating a predictive model using real code. The aim is to demystify deep learning and provide a fast start introduction to those who are new to deep learning.
Intersections of the Marine Corps Planning Process and Data Science
Accurately framing the analytic problem is often the most difficult challenge in the data science pipeline. This talk discusses a potentially helpful framework where a common goal is to improve an organization’s use of data to make better decisions. Application of the framework begins by examining Markov Decision Processes and the agent-environment interface. Key questions related to this interface will drive successful decision-making and provide solutions to guide selection of data sources, algorithms, data visualizations as well as how the organization will use the analytic results.
Data Intelligence: Data-Driven Healthcare
As the state of U.S. healthcare undergoes ongoing changes, the need to implement and execute an effective Data Intelligence program becomes increasingly crucial. Mike Thompson will share his expertise in maximizing data and delivering valuable insights at one of America’s leading hospitals.
The Challenges and Complexities of Retention at Sam’s Club
At Sam’s Club, retention is generally defined in terms of membership renewals – did a member renew at the end of their membership year or not? A multitude of behavioral factors and membership attributes can help us understand the likelihood of a member renewing. However, members don’t always renew on time, some renew early and some renew late. Some members shop very frequently throughout the year and don’t renew while some members don’t shop at all and do. In addition, how to report on renewals is surprisingly challenging.
Networking Coffee Break
Predictive Analytics with Big Data in Health Care
It is often quoted that the future of predictive analytics is in Big Data Analytics. Big Data is seen as a game changer in the retail sector. In health care, Big Data is pivotal in providing proactive care and personalized medicine. In this presentation, Kenny will look at how Big Data is used in the retail and health care sectors. He will illustrate the dark sides to big data, and share best practices on incorporating human intuition into machine learning.
Predictive Analytics: Developing Service Recommendation Systems
Over the last 10 years, Chegg has evolved from a retail company delivering low cost text book rentals, to a major brand in ed-tech. Several of our business lines now provide services in addition to rental services and static content. Expertise in predictive analytics around P2P educational experiences, is something that we have had to develop to maintain product differentiation while scaling. We see service recommendation systems as the next evolution of content recommendation systems (think traditional search). In this talk, we will discuss some of our experiences and learnings.
Navigating through Machine Learning Challenges
Video Games sit at the heart of Gaming Industry and predicting sales of each game is a key input for multiple business decisions at PlayStation. These include fiscal year planning, budget allocation, creation of marketing programs and partnership evaluations. In this presentation, Ankur synthesizes the nuances, challenges and learning from his journey of creating the game sales forecasting model.
Networking Drinks Reception
Registration & Breakfast
An EPiC adventure infusing analytics into Cottages.com
Cottage getaways to the beautiful English countryside have been popular in Britain for many decades, and there is a mature cottage-letting industry in place to serve cottage owners and holiday-makers. Cottages.com is one of the leaders in this space with over 30 years of experience and a portfolio of over 21,000 properties. As with many mature industries, analytics has not traditionally been top-of-mind within these businesses. Come hear about how Cottages.com transformed business processes to be smarter with analytics and in the span of just three years became a truly analytically-led company.
Financial Time Series Forecasting Using Recurrent Neural Network
Time series data is ubiquitous in finance: weekly initial unemployment claims, daily term structure of interest rates, tick-level stock prices, weekly company sales, daily foot traffic recorded by mobile devices, and google search terms, just to name a few. In this presentation, Jeffrey will formulate financial time series forecasting problems using Recurrent Neural Network (RNN) and Vector Autoregressive (VAR) Models, demonstrate how to them implement them using python, and compares their advantages and disadvantages. Given the resurgence of neural network-based techniques in recent years, it is important for finance practitioners to understand how to apply these techniques and the tradeoffs between neural network-based and traditional statistical methods.
Networking Coffee Break
Analytics Powered by Machine Learning
As we are getting into 2018, powered by machine learning permeates everything we do. Predicting price changes from Inventory feeds, improving site navigation through topic modeling, dynamically repricing dealers based on predicted lead close rate, Creating high value audiences in GA for higher ROI etc are all being powered by machine learning. This is a shift that is defining what analytics is responsible for NOW, not in the future. With the advent of tools like SageMaker and CloudML machine learning is no longer in the hands of a select few data scientists, every analyst with access to data can now train models and build production systems that use machine learning. While business problems maybe the same, the the solutions to them are not the same anymore. In order to stay relevant and to build world class products we must embrace this way of thinking.
Data Driven Cities
Cities across the nation and around the world have set up cross functional in house consulting teams called Innovation Teams to tackle major local issues. These teams are set up in partnership with Bloomberg Philanthropies and consist of project managers, designers, and data scientists/analysts. Learn how these teams are drivers of good governance, and specifically how the Los Angeles Innovation Team, empowered by Mayor Eric Garcetti, is using data to solve issues around police hiring and homelessness.
Who Thought Binary Classification Would Be So Complicated!?
As a data scientist on the project management team, I have the responsibility of mapping out the behavior of our players with respect to different features and designs of our game. CCG mobile games are specially designed to be complex and more often than not, the relationship between in-game features and players behaviors cannot be described in a linear fashion. Given the large number of players and the complexity of modeling this interaction, machine learning can play a big role in not only making sense of such high-dimensional data but also in predicting any future value of interest. In this talk, I will discuss one use-case of this approach by defining the problem of imbalanced data sets and introducing 3 effective solutions to tackle them.
P’s of Data Science: Planning Collaborations to Create Products from Data
Our lives as well as any field of business and society are continuously transformed by our ability to collect meaningful data in a systematic fashion and turn that into value. The opportunities created by this change comes with challenges that not only push for new and innovative data management and analytical methods, but also translating these new methods to impactful applications and generating valuable products from data. In a multi-disciplinary data science team, focusing on collaboration and communication from the beginning of any activity improves the ability of the team to bring together the best of their knowledge in a variety of field including business, statistics, data management, programming, and computing is vital for impactful solutions. This talk will overview how focusing on some P’s in the planning phases of a data science activity and creating a measurable process that spans multiple perspectives and success metrics can lead to a more effective solution.
Mining Data to Optimize Design: Analytics for Targeted Affect and User Behavior
Designing digital experiences that support desirable user behavior can be critical to success, especially for businesses with digital products or client-facing portals. This becomes especially challenging when the desired behavior is a hard-to-measure construct (like affect and moment-to-moment engagement) and growth-based behaviors (e.g. learning over time). Toward this end, multi-modal data mining (from structure discovery to prediction and beyond) can uncover emergent behavioral patterns to inform iterative, data-driven design for optimized user experience. This talk presents an ontology of these methods, leveraged during various development stages, for critical insights into key interactions and user behaviors for iterative, informed digital design.
Joint Presentation: How Can We Sell More Vehicles? Leveraging Predictive Analytics to drive customer retention and sell more cars.
Presented by: Jeremy Coltin, Director, Dealer Programs & Predictive Analytics, Hyundai Capital AmericaPeter Kim, Senior Manager, Predictive Analytics, Hyundai Capital AmericaAs the automotive industry continues to look to the future; most organizations are stuck in the present, grinding month-to-month in order to hit sales objectives and retain customers. What happens when you insert predictive analytics into the inner workings of the old fashioned car dealership? See how Hyundai Capital America has become a driving force in marketing intelligence through the use of predictive analytics to reach the right customer at the right time.
Networking Coffee Break
Applications of Predictive Analytics in Marketplace Lending
Marketplace lending or online lending is a growing and evolving space that is heavily dependent on data and predictive analytics. It leverages data from both sides of the marketplace – borrower data as well as market / investor data – to offer better rates to borrowers and attractive returns to investors. Predictive analytics is fundamental to the overall credit underwiring and risk management framework, including borrower marketing, credit decisioning / pricing, verification and servicing. As this ecosystem continues to evolve, leveraging new data sources to build better predictive models and reducing friction in the process is important for the success of the platforms.
Forecast of Financial Measures from Opinions using Machine Learning
The convergence of recent data and machine learning algorithms shows the potential power to replace or augment the predictive power of traditional analytics. We look at a few methodologies and at the increasing importance of clean data in the financial domain.