Category Archives: predictive analytics

SQL Saturday #279 in Phoenix – March 29, 2014

If you haven’t noticed, SQL Saturday #279 in Phoenix this weekend has 3 great BI pre-cons that are scheduled for Friday, March 28.  The pre-cons are held at the Microsoft Office in Tempe, AZ and run from 8:30AM – 4:00PM. There is still time to register!

I’ll be presenting Predictive Analytics in the Enterprise, showing you how to start and be successful with a predictive modeling project in your organization.  I’ll provide an overview of predictive analytics, discuss the concepts and ideas behind the subject, use case studies to review successful projects I’ve worked on, and demonstrate how predictive models are built.   I’ll also have a number of labs for the class to use as exercises so you can leave the session having built your own predictive models.  To participate, bring your laptop or plan to share with someone.  I have a lot of material to go through in just one day.

The other two pre-cons will be led by David Stein and Bill Pearson and are listed below.  Also, don’t foget Saturday, a full day of free training.  I’ve spoken at this event in 2012 and 2013, and it’s a top notch gathering of SQL Professionals.  I expect it to be this year also.

Session 1: Fast Track your Business Intelligence Projects with BIML – David Stein

Session 2: Predictive Analytics in the Enterprise – Carlos Bossy

Session 3: Practical self-Service BI with PowerPivot for Excel – Bill Pearson

BI and Analytics in 2014

It’s become a tradition at the end of the year for bloggers and industry pundits to make predictions about their fields of interest, and business intelligence is not spared this annual ritual. While I’ve avoided (or forgotten) making my own predictions, it seems that the future always comes up when I’m speaking at an event. People want to know what’s coming so they can prepare.

Four Areas to Explore

With the PASS BA conference coming up in May in San Jose, the question of future direction will certainly come up again. The field of Business Intelligence is undergoing a shift in how it’s delivered to users and the way it’s developed. After a period of having established common data architectures and practices, including Star Schema models, ETL processes, and cubes for what-if analysis, we are seeing new technologies driving us in different directions. In my opinion BI Architects and Practitioners should start thinking about the following:

1 – Adding predictive analytics to BI teams to get even more value from their data.

2 – Look into data virtualization for some reporting solutions.  This can greatly reduce the time and cost of ETL development.

3 – Consider putting some data in the cloud, when it makes sense.

4 – Make Excel a centerpiece of your designs. If you’re like a lot of BI Architects then you’ve been trying to displace Excel for years, but now with so much BI functionality, it’s worth keeping.

Take the time to learn more about these topics and find out how they fit into you architecture’s future.  These topics will be covered in multiple sessions at the PASS BA Conference.

A full day of Predictive Analytics

The annual SQL Saturday event in Denver is scheduled for this Saturday, September 28, but you may not know that there is also a full day of training just one day earlier.  There are a total of three sessions scheduled for Friday and I have the privilege of presenting one of them, Predictive Analytics in the Enterprise.  If you’ve been reading the tech news and paying attention to where analytics is headed, predictive analytics is a great topic to dive into, especially for BI developers, analysts, architects, and DBAs.  Here’s a look at the session abstract:

Data Science is a red-hot subject, but what does that mean for SQL Server DBAs and developers?  The Data Mining algorithms in Analysis Services provide capabilities to achieve predictive analytics and modeling with your SQL Server data and any other data you have access to.  In this session you’ll get the chance to cover the data mining process from start to end.  By the end of the day, you’ll have a practical understanding of how to develop predictive models for your data.  The session includes a brief overview of data mining followed by an opportunity to complete actual predictive modeling exercises.  You’ll see how to fine-tune and compare the results using various algorithms, and how to make the best use of the models in your enterprise.   With Predictive Analytics having achieved the Plateau of Productivity in Gartner’s Hype Cycle for Emerging Technologies, now is a good time to learn what these powerful tools can do for you.

Like the abstract says, the session will cover the concepts and terminology in the predictive analytics world and will include case studies to review.  It will also include instructor-led labs so you can create your own predictive models.  Bring a laptop with SQL Server Analysis Services (multi-dimensional),  Business Intelligence Development Studio (or SQL Server Data Tools), Excel and the Data Mining add-in for Excel.  If you don’t have a laptop with these tools readily available, we can pair you up with a partner who has the tools ready to go.  Working in pairs will also be a good way to learn the topic better.

If you’re looking for training in other areas, Tim Mitchell will have a full day session for ETL developers on Real World SSIS: A Survival Guide, and Dmitri Korotkevitch will target DBAs with a session titled SQL Server Internals from the Practical Angle

See you Friday!