Book Image

Data Analysis Using SQL and Excel - Second Edition

By : Gordon S. S. Linoff
Book Image

Data Analysis Using SQL and Excel - Second Edition

By: Gordon S. S. Linoff

Overview of this book

Data Analysis Using SQL and Excel, 2nd Edition shows you how to leverage the two most popular tools for data query and analysis—SQL and Excel—to perform sophisticated data analysis without the need for complex and expensive data mining tools. Written by a leading expert on business data mining, this book shows you how to extract useful business information from relational databases. You'll learn the fundamental techniques before moving into the "where" and "why" of each analysis, and then learn how to design and perform these analyses using SQL and Excel. Examples include SQL and Excel code, and the appendix shows how non-standard constructs are implemented in other major databases, including Oracle and IBM DB2/UDB. The companion website includes datasets and Excel spreadsheets, and the book provides hints, warnings, and technical asides to help you every step of the way. Data Analysis Using SQL and Excel, 2nd Edition shows you how to perform a wide range of sophisticated analyses using these simple tools, sparing you the significant expense of proprietary data mining tools like SAS.
Table of Contents (18 chapters)
Free Chapter
1
Foreword
17
EULA

Lookup Model for Probability of Response

This section considers a different sort of problem, related to the subscription data. What is the probability that a customer who starts in 2005 is going to last for one year?

The Overall Probability as a Model

The way to start thinking about this problem is to consider all customers who start in 2004 and ask how many of them survive for exactly one year. Using one year of starts dampens seasonal effects occurring within a year. Also, the subscription table has no stops prior to 2004, limiting how far back in time we can go.

Chapter 8 addressed several different methods for looking at survival and retention. This section looks only at the point estimate after one year, as calculated by the following query:

SELECT AVG(CASE WHEN Tenure < 365 AND StopType IS NOT NULL
                THEN 1.0 ELSE 0 END) as stoprate
FROM Subscribers
WHERE YEAR(StartDate) = 2004

Customers who stop within one year have tenures less than one year and a non-NULL stop...