2014-01-14: Adding Predictive Analytics to BI Team & Random Forest & Decision Trees using R
University of Colorado Denver - Tuesday January 14, 2013 @ 6:15pm MST
Adding Predictive Analytics to Existing BI Team - Abstract
Many organizations have a Business Intelligence group or several that provide reports and data to leadership. A majority of these activities revolve around descriptive analytics trying to answer the question “What has Happened?”. Our presentation focuses on the challenges and lessons learned of adding improved analytics and predictive analytics
skills to an existing BI team without breaking the budget. We discuss our grassroots approach and lessons learned through the early stages of our journey to provide greater insights to business decision makers.
Mike Little: Mike is Vice President of Management Reporting in Finance at Level 3 Communications. He is responsible for global financial business intelligence, predictive analytics, financial data governance, sales productivity reporting/analysis, and sales commissions. He has been with the company for over 11 years and has held numerous business intelligence / software development positions in finance, sales operations, and IT. Prior to Level 3, he worked as a consultant specializing in web application development. From a big data/predictive analytics perspective, they support multiple areas of the business by focusing on customer churn, employee churn, forecasting, and cross-sell/up sell opportunities. He has a BS with an emphasis in Information Systems and Human Resource Management from the University of Colorado, Boulder.
Greg Arnold: Greg Arnold is Senior Manager of Data Science in the finance organization at Level 3 Communications. He has been with the company for 14 years in several roles from operations to business process engineering and IT. Prior to joining Level 3 he served in the US Navy as an Electronics Technician. He has a BS in Information Technology with an emphasis in web design from Kaplan University.
Random Forest and Cart Decision Trees using R - Abstract
Data Science using open source tools, building and evaluating models, using decision trees and ensemble methods such as random forest. This talk will cover the importance of data preparation and discuss a few algorithm implementations. Data science problems are not solved by just throwing together complex algorithms, advanced machinery, time, and data (and we already know the answer to that is 42). It takes creativity to prepare the analytic data sets, and the best algorithm to use depends on the data and the business question.
Jennifer Evans is a Data Scientist at Clickfox. She went to the Colorado school of Mines, studying Computer Science, Applied Mathematics and Engineering Physics. Clickfox is the leader in costumer experience journey analytics. Taking data from disparate sources, tying it together, and creating a unique individual costumer journeys across channel. The data science team provides analytics of a whole new kind of data at scale, with focus on building a software platform. The CEA Platform maps data across interaction channels and identifies the journeys that matter for the optimal decision-making on key business drivers.
Register @ http://bit.ly/1jbHmBp