Data Science: Why? & Cubes in Production

When you registered, you asked to be notified about new content. Read on, or scroll to the bottom for a link to unsubscribe instantly, promise! 

New

What is Advanced Analytics, Data Science, Machine Learning—and What is their Value?

The most common question I get asked by budding data scientists is: how can I explain the value of data science to my customers?  I wrote an in-depth article that explains the five key reasons for doing advanced analytics in terms of the short-term and strategic value it provides to business customers. It summarises almost ten years of my project experience in this area, and I share some observations about the successes and failures I had, which you won't easily find shared anywhere else. I hope that it can save you some trouble, and perhaps help you justify doing data science. At the same time, I explain some of the key terminology, covering the differences between data mining and machine learning, what makes data science into a science, and how does it differ from advanced analytics.

Above all, I hope that you will want to understand and apply the scientific method of reasoning to your own data-driven quests for better decision making. Indeed, it is that very principle, based on the idea of testable hypotheses, experiments, and a small dose of statistics and machine learning, that enables us to have more confidence in data. To a sceptic, even if nothing has changed, everything changes, this way, and I have now lived through three hype-disilusionment cycles surrounding machine learning over the last 30 years. I am sure we will find much disillusionment once again: I already see people building and speedily deploying unvalidated or downright invalid models. Fortunately, those who are ready will benefit handsomely from the "new" data science. I hope you enjoy reading my thoughts, and please share your feedback.

Practical Data Science with Azure ML, R and SQL

I wrote, and I teach, what is still the first-in-the-world, hands-on, intensive 4-day classroom course on Practical Data Science with Cortana Analytics: Azure Machine Learning, SQL Data Mining and R. I have delivered it already in Dublin, and to sold-out rooms in Stockholm (twice) and London. If you are in North America, you can still book the one last remaining place (as of today) on the course in Chicago (16 Nov). We have a few spaces left in Zurich (23 Nov), and Oslo (30 Nov), and good availability in just-added Copenhagen (11 Jan). Click here to read the detailed syllabus, to find out the dates, and to book your place.

You are welcome to bring your own data, to get a chance to analyse it with my help during the course, or use my training data to solve the data science puzzles I will task you with during the course. See you in the real world!

SSAS Cubes: Security, Deployment, Performance with Chris Webb

The final instalment in Chris Webb's popular course on SSAS Cube Design teaches you how to make your cubes fast and effective, by designing appropriate aggregations for them, a subject I am sure any cube-aficionados will cherish to study, especially from none other but the very world-expert on this matter. You will also find out how to balance performance goals achieved while querying with any degradation of cube processing time. As usual, you will see plenty of detailed demos, even focusing on the minutiae of such aggregation design concerns as many-to-many relationships, semi-additive measures, partitioning strategies, parent-child hierarchies and MDX calculations.

The remaining two new cube design modules cover its operational aspects. Cube Deployment, Processing and Admin also discusses backup and restore, synchronisation, and scripting. Cube Security teaches you about administrative and data security, the latter on three levels: at cube, dimension hierarchy, and at cell levels. I am sorry to say that this completes Chris's SSAS cube design course. While you enjoy his latest work, I already look forward to working with Chris on his future content.

Newsletter Promo: $40 Off

Following the announcement I made in June, our pricing has changed. We offer annual ($290) and 6-month ($190) Full Access Memberships to individuals and organisations as of this week. Organisations can benefit from a discount built-into our economical group memberships. And all the newsletter readers can take $40 off when subscribing to, or renewing any memberships, by applying the following code, valid until the end of November:

USD40NEWSLETTER2015NOV

I wish to keep the price of our training as low as possible. However, I also need to keep our authors happily dedicated to you. We will be changing our prices once again in early 2016, moving to a per-course pricing model. If you renew now, your new membership will extend your existing one, and nothing will change for you while it remains valid, even after we have introduced the new pricing—as long as your membership has not expired.

I believe the new structure will let us provide you with great content for years to come. If you have any suggestions, or questions, please get in touch, I always look forward to your feedback.

If you happen to have any unredeemed 1 or 3 month group memberships, please get in touch to use them, as they cannot be redeemed online at the moment. Other prepaid memberships and discounts work online as usual.

Enjoy learning data science, advanced analytics, and BI with us.

Rafal