Online Program

Return to main conference page
Keynote Address | Concurrent Sessions | Poster Sessions
Short Courses (full day) | Short Courses (half day) | Tutorials | Practical Computing Demonstrations | Closing General Session with Refreshments

Last Name:

Abstract Keyword:



Viewing Practical Computing Demos onlyView Full Program
Saturday, February 17
PCD1 Deploying Quantitative Models as 'Visuals' in Popular Data Visualization Platforms
Sat, Feb 17, 2:00 PM - 4:00 PM
Salon E
Instructor(s): Daniel Fylstra, Frontline Systems Inc.
Data visualization and business intelligence tools such as Tableau and Power BI have become extremely popular in recent years. Tableau reports that over 90% of Fortune 500 companies are now customers, while Microsoft reports that over 200,000 organizations of all sizes are using Power BI. These tools currently offer easy-to-use access to many data sources, powerful facilities for "slicing and dicing" data, and rich, flexible data visualization, but only limited built-in analytics methods.

A new avenue has emerged in the past year for extending analytics methods in both Tableau and Power BI- and this provides a new way for an analyst to develop quantitative models outside these platforms, then deploy them as 'visuals' inside Tableau and Power BI, in 'dashboards' which are often published for use by thousands of users in an organization. Though originally conceived as a way to extend the range of visualization styles, these components can perform arbitrary computations on data before it is rendered in visual form.

In this session, Excel Solver developer Frontline Systems, one of the first to explore this new avenue, will demonstrate use of its tools to automatically convert existing quantitative models into 'visuals' for both Tableau and Power BI. Among other options, this enables an analyst to convert predictive (data mining, machine learning) or prescriptive (optimization, simulation) model from Microsoft Excel into an easily-deployed 'visual', just two mouse clicks. No programming is required, but the ability to extend models using high-level RASON modeling language code or programming language code is available. These 'visuals' are full-fledged models that easily connect to any Tableau or Power BI data source, and re-solve the underlying problem whenever the data sources are refreshed.

PCD2 Handling Missing Data Using Multiple Imputation
Sat, Feb 17, 2:00 PM - 4:00 PM
Salons BC
Instructor(s): Yulia Marchenko, StataCorp LLC
This workshop will cover the use of Stata to perform multiple-imputation analysis. Multiple imputation (MI) is a simulation-based technique for handling missing data. The course will provide a brief introduction to multiple imputation and will demonstrate how to perform multiple imputation in Stata. The three stages of MI (imputation, completed-data analysis, and pooling) will be discussed with accompanying Stata examples. Imputation using multivariate normal (MVN) and using chained equations (MICE, FCS) will be discussed. A number of examples demonstrating hot to efficiently manage multiply imputed data within Stata will also be provided. Linear and logistic regression analysis of multiply imputed data as well as several postestimation features will be presented. No prior knowledge of Stata is required, but basic familiarity with multiple imputation will prove useful.