Data Mining: Model Building, Testing & Predicting

When you registered on ProjectBotticelli.com, you asked to be notified about new content. This is our sixth newsletter since we have launched our site.

Summary of New Content:

• Data Mining Model Building, Testing and Predicting with Microsoft SQL Server and Excel (80 min)
• HappyCars Sample Data Set for Learning Data Mining (database + SSDT project)
• Final PPTs from BI and Productivity Roadshow 2012, covering SQL Server 2012, Power View, and Big Data

News

You will be hearing a lot about Business Analytics over the next year, and beyond, as the domains of Business Intelligence, Predictive Analytics, Big Data Analytics, Data Warehousing, ETL and Data Integration are all slowly merging together, while exchanging and cross-breeding the underlying technologies. Data Mining, which combines statistics, machine learning, and database technologies, is only a few decades old—except for Bayesian algorithms, which were first thought of in 1700s. Big Data Analytics is relatively new, introducing large-scale, inexpensive processing parallelism to analytics, primarily by means of Apache Hadoop. In the Microsoft world, Hadoop is now well integrated, and you will find other forms of parallelism in the world of data, for example in the Microsoft Parallel Data Warehouse.

It is clear, that all of those approaches share the same goals: to provide you with meaningful answers about knowledge that has been hiding in your data, be it: small or large, slow or fast, simple or complex, structured or not. Whilst Data Mining is relatively mature, and well past the famous Gartner Trough of Disillusionment phase, Big Data is still on the way up the hype slope. However, a few years from now, we will no longer need to distinguish between the individual underlying ideas, various learning algorithms, ability to compute more in-memory, as opposed to scanning mechanical or SSD disks, or choose between relational, multidimensional, tabular, or unstructured data. All of those building blocks of the future of Business Analytics are about to be integrated, perhaps giving rise to, simply, Analytics, without any adjectives, modern enough by 2020 standards. Incidentally, if you are interested in my thoughts on the naming of the things we do, have a look at my older article about Performance Intelligence, some time.

In the meantime, Data Mining is undergoing a revival of interest, and now is a good time to learn it, if you are new to it, or to practice it more, if you have been doing it for a while. That is the key reason why we have been focusing our efforts on helping you learn that little-known, mature, and powerful, component of SQL Server Analysis Services. Our new Data Mining series is a great place to study it. It already includes 3 modules, and a sample data set, to help you practice it, and more modules are planned. The first module is free for everyone to watch, the remaining ones are for Full Access Members:

• Introduction to Data Mining with Microsoft SQL Server (24 min, New!)
• Data Mining Concepts and Tools (50 min)
• Data Mining Model Building, Testing and Predicting with Microsoft SQL Server and Excel (1 h 20 min)
• HappyCars Sample Data Set for Learning Data Mining (database + SSDT project, New!)

You can find out more about the newest content further below.

An Apology, and a Special Offer—10% Discount

Some of you, who have a VAT registration number in EU, have experienced difficulties renewing your Full Access Membership using a discount code. I would like to apologise for that. I'm pleased to say that the issue seems to have been resolved—thanks to Sanna from Finland, for alerting us to it. To make up, I would like to extend the 10% discount to all of you, both renewing or just signing up, valid on all memberships, until the end of July 2012—use coupon code JUL2012NEWSLETTER at checkout. Feel free to share it, or tweet it.

Group Membership

If you work for a company, Data Mining training could be useful to your colleagues. You can get Full Access Membership for multiple users by purchasing a discounted Group Membership pack. It is very simple: buy a pack, receive pre-paid membership codes for your users, then each user redeems a code. You also get one tax invoice for the whole purchase and everyone looks after their own usernames, passwords, and preferences. Above all, buying a Group Membership is cheaper than buying individually, and—until the end of July—you can use that 10% off code, mentioned above.

New Content in More Detail