Better Software Through Measurement

This course will show you how to generate recommendations for your users, filter messages based on users' preferences, decide which web page performs best, keep track of timings in your application, and discover groups among items.
Course info
Rating
(32)
Level
Intermediate
Updated
Oct 10, 2013
Duration
1h 37m
Table of contents
Instrumentation: Streaming Metrics
Optimizing Conversion: A/B Testing
Recommendations: Item-based Recommendations
Personalized Recommendations: Naive Bayesian Classifier
Finding Groups: k-means Clustering
Description
Course info
Rating
(32)
Level
Intermediate
Updated
Oct 10, 2013
Duration
1h 37m
Description

This course will show you how to generate recommendations for your users, filter messages based on users' preferences, decide which web page performs best, keep track of timings in your application, and discover groups among items. These techniques are at the heart of many of the largest search engines and online retailers, but can be used to good effect for smaller companies. Throughout the course, the emphasis will be on examining and extending working sample code. The algorithms will be presented intuitively and you do not need any advanced math background.

About the author
About the author

Brendan started programming in C on MacOS 7.5 and since then has done everything from embedded systems programming to writing web apps.

Section Introduction Transcripts
Section Introduction Transcripts

Optimizing Conversion: A/B Testing
Hi, this is Brendan Younger, and welcome to this module on optimizing website conversion rates using A/B testing. In this module, I'll show you how to run experiments on your website, to determine how to best achieve your goals. Whether this means email signups or completed sales or anything of the sort.

Recommendations: Item-based Recommendations
Hi, this is Brendan Younger and welcome to this module on making recommendations to your users. Recommending additional products or services that a user might like is one of the easiest ways to increase your revenue and keep your users coming back for feature purchases. In this module, I'll show you how to create a simple yet effective recommendation engine that takes into account other user's buying habits to find additional products to recommend.

Personalized Recommendations: Naive Bayesian Classifier
Hi, this is Brendan Younger and welcome to another module on making recommendations to your users. In the previous module we saw how to create a recommendation engine that took into account other users actions. Now that works well when you have many users and relatively few items to recommend. However, sometimes users have very personalized information presented to them. And you want to use their unique preferences to sort that information for them. Now examples of this would include personalized email inboxes and news feeds.

Finding Groups: k-means Clustering
This is Brendan Younger, and welcome to this module on Finding groups within your data. There is a natural human tendency to try to find commonalities between data and discover groups. You hear this in the media when they decide that all people of a certain generation are behaving the same way, and therefore should be called generation x or generation y. Well, as fun as it is to come up with groupings on your own, there are several algorithms which are capable of clustering data together, and perhaps finding groupings that you didn't know were already there. In this module, we'll look at one such module, known as K-Means clustering.