Community Spring Cleaning week is here! Join your fellow Maveryx in digging through your old posts and marking comments on them as solved. Learn more here!

Analytics

News, events, thought leadership and more.
DrDan
Alteryx Alumni (Retired)

We are excited to announce a brand new district on the Alteryx Analytics Gallery, the Predictive District. This district is now available to the public and contains new tools, usage samples, and analytic apps designed to enhance the predictive analytics capabilities of Alteryx, and make advanced analytics more accessible to data analysts with easy apps and examples to quickly leverage some of the predictive tools within Alteryx. The Predictive District will continue to grow over time through a combination of new capabilities created by the Alteryx product team, and through contributions from the Alteryx user community.

 

Introducing the Predictive DistrictTo get things started, we are launching eight new predictive tools and five new usage samples in the Predictive District. The new tools introduce new advanced analytics methods which assist in model diagnostics and comparison, offer the ability to export models created in Alteryx to external scoring engines, and a tool that extracts estimated coefficients from models created using the Count, Gamma, Linear, and Logistic Regression tools into an Alteryx table (as opposed to a report element) that can be used as input into other Alteryx tools and calculations.

 

All of the tools within the Predictive District stem from customer feature requests, and several of them have been around for some time, having been created to address the needs of specific customers, but are useful to a broader set of users.

 

The new tools being released in the Predictive District include:

 

  • Model Coefficients: This tool takes as input a model created by the Count, Gamma, Linear, or Logistic Regression tools and provides a data table of the estimated model coefficients that can be used as input into other, downstream, Alteryx tools. While some users will find this capability useful, we believe that one of the key benefits of this tool is to act as a simple to follow template for users to develop their own macro based tools for extracting information from an R model object within Alteryx.
  • Model Comparison: A tool that allows the user to compare the predictive efficacy of different models in predicting a target variable of interest. This tool is applicable for comparing models for both continuous and categorical target variables, providing the appropriate set of comparison measures given the target's data type.
  • Model Output: A tool that allows a predictive model created in Alteryx to be written to a file in either R's native binary file format or PMML, allowing the model to be used in an external scoring application.
  • Multidimensional Scaling: A method for identifying underlying dimensions using the actual or perceived distance or dissimilarity between items, which is often used to identify perceptual dimensions in marketing research applications.
  • Importance Weights: Bivariate measures of the association between a target variable of interest and one or more potential predictor variables. Both the target and the predictors can be either numeric or categorical. This tool allows a user to quickly determine which variables, out of a large set of potential predictors, will be effective in predicting a target variable of interest.
  • Survival Analysis: Core survival analysis routines, which are frequently used in churn analysis applications. The goal of survival analysis is to determine how likely some event is going to occur over a specified period of time.
  • Survival Score: This tool provides predicted score measures for Cox proportional hazards models that are appropriate for survival analysis models.
  • Variance Inflation Factors: A variance inflation factor is a measure of how much the standard error of an estimated model coefficient for a predictor variable has been increased as a result of the correlation between that variable and the other predictor variables, which can provide important diagnostic information in the model building process. The output of this tool has been designed to allow an analysis of variance inflation factors to be added to the standard reports generated by the Count, Gamma, Linear, and Logistic Regression tools.

 In addition to the tools, five samples are included in the initial wave of items in the Predictive District. The samples are:

 

  • MB Affinity which demonstrates the use of the existing MB Affinity tool in Alteryx.
  • Model Comparison Sample which demonstrates the use of the Model Comparison tool.
  • Multidimensional Scaling which demonstrates the use of the new Multidimensional Scaling macro.
  • Survival Analysis Sample which demonstrates the use of the new Survival Analysis and Survival Score tools.
  • Variance Inflation Factors and Model Coefficients which demonstrates the use of these two new macros.

 

This is the first wave in what will be an expanding set of predictive oriented tools, macros, and workflows. We already have a number of new tools in the works, and encourage others to join in adding to the expanding list of predictive analytic samples, workflows and apps in the Alteryx Gallery by making contributions to the Predictive District. We believe that using the Predictive District as a channel for distributing new Alteryx functionality will better allow us to provide users of all skill levels the functionality they need to address the analytical challenges they face, without overburdening their tool palettes, make predictive analytics more accessible, and reduce the need for advanced users to write their own custom R code in Alteryx since they can make use of new functionality in the Predictive District.

 

We encourage you to visit the new Predictive District and experience all of the new samples and tools available to you.

Dan Putler
Chief Scientist

Dr. Dan Putler is the Chief Scientist at Alteryx, where he is responsible for developing and implementing the product road map for predictive analytics. He has over 30 years of experience in developing predictive analytics models for companies and organizations that cover a large number of industry verticals, ranging from the performing arts to B2B financial services. He is co-author of the book, “Customer and Business Analytics: Applied Data Mining for Business Decision Making Using R”, which is published by Chapman and Hall/CRC Press. Prior to joining Alteryx, Dan was a professor of marketing and marketing research at the University of British Columbia's Sauder School of Business and Purdue University’s Krannert School of Management.

Dr. Dan Putler is the Chief Scientist at Alteryx, where he is responsible for developing and implementing the product road map for predictive analytics. He has over 30 years of experience in developing predictive analytics models for companies and organizations that cover a large number of industry verticals, ranging from the performing arts to B2B financial services. He is co-author of the book, “Customer and Business Analytics: Applied Data Mining for Business Decision Making Using R”, which is published by Chapman and Hall/CRC Press. Prior to joining Alteryx, Dan was a professor of marketing and marketing research at the University of British Columbia's Sauder School of Business and Purdue University’s Krannert School of Management.

Comments