All About Squark
Learn how to use Squark from start to productive implementation of results.
Introduction to Squark
Learn what Squark automated machine learning for marketing is all about—creating actionable predictions.
What Model Classes Does Squark Support?
Learn about binary (yes-no), multivariate (which of three or more), and regression (forecast of numerical value) models.
What AI Algorithms Does Squark Use?
An overview of the AI algorithms used automatically by Squark to find the most predictive model for your data.
What Are Training and Production Data?
Why you need 1. a training data set (from which Squark learns patterns), and 2. a production data set (containing the records for which Squark will add predictions in the target variable column(s)).
Loading Data Into Squark
How to load training and production data into Squark.
What is Feature Selection?
Explanation of selecting columns (features) for modeling and prediction in Squark.
Feature Engineering - Imputing Data
Learn how Squark automatically adds detail to data sets where possible in order to increase their predictive value.
Text Vectoring - How Word2Vec Makes Text Strings Predictive
Learn more about Natural Language Processing (NLP) in Squark that can make text strings predictive.
Understanding Squark Results
See how Squark results can be understood and put into action.
Leaderboard Sorting
How does Squark determine the rank of each model on the leaderboard for our three different types of analysis?
Variable Importance Explained
See how Squark reveals which data features (variables) are most predictive so you can inform action plans accordingly.
Understanding the Confusion Matrix
What does the Confusion Matrix show on the result screen when you run a binary or multinomial classification?
How to Predict Churn - a Binary Classification (Yes-No) Example
Here is an example of predicting the likelihood of an outcome for each record where there are only two possible answers. Cable subscription data is used. Find the sample churn data sets here for your own practice.
How to Predict Price - a Regression (Forecast) Example
For pricing predictions, Squark automatically uses algorithms the are best for regression, or forecasting numerical values such as price, in this example. Learn how to do it. Download sample pricing datasets here for your own practice.
How to Predict Coupon Offer Response - a Multivariate (Multiclass) Classification Example
In this example you will learn how to predict the likelihood of offer success when there are three or more possibilities. Download sample datasets for coupon multivariate classification for your own practice
System Overview and Deep Dive
Learn more about how Squark works underneath its deceptively simple user interface.
Automated Machine Learning
Learn more about the benefits of automated machine learning as well as the AI power underpinning Squark.
Eric Kavanagh of The Bloor Group Introduces ‘Actions Beat Insights - How AutoML Can Expedite Business Value’ Webinar
Actions Beat Insights Part 1 - Introduction to AutoML
Actions Beat Insights Part 2 - Model Types: Binary, Multi-class, Regression, Time Series
Actions Beat Insights Part 3 - The AutoML Process with Use Cases
Actions Beat Insights Part 4 - Q and A
Squark Use Cases
Learn more about using Squark to solve previously unsolvable problems.
Nielsen B2B Marketing Use Case
How to Predict Churn - a Binary Classification (Yes-No) Example
Here is an example of predicting the likelihood of an outcome for each record where there are only two possible answers. Cable subscription data is used. Find the sample churn data sets here for your own practice.
How to Predict Price - a Regression (Forecast) Example
For pricing predictions, Squark automatically uses algorithms the are best for regression, or forecasting numerical values such as price, in this example. Learn how to do it. Download sample pricing datasets here for your own practice.
How to Predict Coupon Offer Response - a Multivariate (Multiclass)