Most businesses, including medical clinics, want some method for forecasting business trends. The Cognos BI powerplay presentation tools contain the tools needed for basic forecasting.  The trick in getting accurate forecasting is select the correct algorithm.

I recommend trying the model against only a portion of your data, to see how closely forecasts the remaining data.  For example, if you have 2 years worth of data, try entering only 18 months of data into the model and let it predict the final 6 months.  Then you have some actual data with which to measure ‘fit’ of the prediction.  The new Microsoft SQL tools mentioned below, allow training data and measuring fitness.

Forecasting SSAS vs Cognos

IBM Cognos Forecast Models:

The old Cognos website used to list more details about each model, but that has long since been eaten by IBM’s user un-friendly website.  I’m sure a few of my friends still enjoy ‘Recreational Math’ magazine, so here are some algorithms to play with in your spare time.

For more details,  read about the terms of service and limitations of liability on their forecasting algorithms pages.

Microsoft SQL Analysis Services

If you are using the Microsoft Business Intelligence stack, there are several data mining options in Analysis Services (SSAS).  The SQL team has worked hard to lower the barriers to entry and learning curves for data mining.   The Excel add-in makes the SSAS data mining models easily accessible to business users at the desktop.  I will have more about these tools soon.

Again, for details and limitations, see the Microsoft articles on each algorithm.

Just for a quick overview, in Visual Studio for 2012 (BIDS for previous versions) your project properties deployment will point at an SSAS Server Instance which in turn has to have query logging configured to point to a SQL Server database.  That database table, then logs info about the queries including the name of the server, SSAS database, cube and measure group.  So it really forms a circular reference pointing all the way back around to the project in which you are working.

SSAS Usage-Based Optimizer Connects from the SSAS Project, to the server, gets the configurations and connects to SQL Server logging database. The names have to match all the way around the circuit.

There are a few basic things to check:
1. Check that the SSAS Database name equals the Object ID. If you have renamed the .cube file in the SSAS project, then these can get out of sync.
2. Check connection strings on your SSAS server configuration page.
3. Check SQL permissions on the database where SSAS is trying to log the queries.
4. Query the log table to make sure it has results which match the cube and measure group.
5. Fix the log table entries if you have moved the logs from Production down to Development.

Moving Between Environments:

If you log queries in Production and move the log to Development, you will have to run some update statements to fix your log table.   The MSOLAP_ObjectPath field contains a fully qualified reference to the SSAS server, database, ObjectID (Cube) and Measure Group.

  MSOLAP_ObjectPath explanation:
  SERVER   .Database                      .ObjectID          .MeasureGroupName
  COGNITIVE.Analysis Services Tutorial L10.Adventure Works DW.Internet Sales

To change server name references in the log table, just run a quick update script:  
  USE SSASLogsCognitive 
  GO
  UPDATE dbo.OlapQueryLog
  SET MSOLAP_ObjectPath = REPLACE (MSOLAP_ObjectPath, 'COGNITIVEPROD.','COGNITIVEDEV')
  GO

Common Error Messages:
“No Query Results…” 

If the table is empty, check permissions to make sure the SSAS Service account can write to that database.

If the table has results, check the MSOLAP_ObjectPath and be sure the measure group has some results.

“Cannot connect to database… Host actively refused connection”

Check the connections listed above, starting from your Project to your SSAS Server to SQL.

References:

http://www.jenstirrup.com/2009/04/optimising-aggregations-in-analysis.html

http://www.sqlchick.com/entries/2012/5/5/why-is-my-ssas-query-log-table-empty.html

Handy Tool for Aggregations:

http://bidshelper.codeplex.com/wikipage?title=Aggregation%20Manager&referringTitle=Documentation

EntreLeadership tour in Oklahoma City, OK. 4-19-2013

Last night a friend called and invited me to attend Dave Ramsey’s EntreLeadership 1 Day event.  Being an entrepreneur, I’m always interested in learning more about business.  I’ve listened to the EntreLeadership audio book, so I just hoped to learn more of the same.  The best surprise of the day was the hour by Jon Acuff about his new book START, Punch Fear in the Face, Escape Average and Do Work that Matters.

For about an hour Jon talked about the phases we go through in business and life.

The Road to Average

These are the ‘typical’ stages along the path most people take through life.

In our 20′s, we are Learning.  We are leearning that the only thing we really control is the starting line.  The finish line may move or change along the way.  He also encouraged people to start before they are ready because there never is a golden moment of ‘ready’.

In our 30′s we are Editing.  I enjoyed Jon talking about editing out the things you don’t enjoy and doing the things you do enjoy.

In our 40′s we are Mastering.  I’m here now and enjoying it.  I’ve spent over ten years doing data warehousing and business intelligence systems.  I’ve learned the hard way and edited out what didn’t work.

In our 50′s we are Harvesting.  One way or another, we harvest what we sow.  Very biblical principle at work there.

In our 60′s we are Guiding.  I hope by taking time to teach and share, I’ve been guiding along the way.

The Road to Awesome

What struck me most was Jon talking about the ‘back road’, which is less traveled. There is a generation that has grown up expecting things to be readily available online and they cycle through the same phases faster.

He also talked about not listing to the ‘voice of fear’.  My favorite quote from Jon was “Fear only bothers you on things that matter.”   So true.

I picked up a copy of his book and I’m reading it this weekend.

Get a copy.  Be Awesome!

This is one of the most fun and insightful business books that I’ve read in quite a while. It is parable about four behavioral styles (DISC profiles) that affect every interaction around us.
Set in a forest called Home, there is Dorian the Dominant eagle who is doesn’t chit-chat or pull any punches.
The fun loving parrots like Indy are Influencers who just wants to party. The doves like Sarah enjoy the Steady life of Home and want everyone to get along. Finally, the owls like Crystal are very Compliant and need everything to be planned out and perfect. Sound like anyone you know?

This is a fun evening read that will give practical advice on dealing with, you know, humans…

Sharing SQL data in SharePoint

In March, I was invited to speak at the OKC SQL Users Group and I never got the slide deck posted. So here it is to help get you started.

Session Level:  Beginner SQL and Beginner SharePoint

Session Abstract:  Microsoft SharePoint is a collaboration software that provides a web based portal for business intelligence, content management and document sharing.

• In this session we will start with the basics of sharing a SQL table as a ‘List’ in Sharepoint.
• Next we move on to fun stuff of integrating SQL Server Reporting Services integration.
• Finally we share a PowerPivot analysis built off our SQL data.

For those who asked about the cool Excel PowerPivot charts, go check out the articles and book from Roggie Clark, The SQL Surgeon.

Please be kind with your comments, as this is just the first time I presented on SharePoint. I got great feedback and will continue to add to the session.

Special thanks to our friends at GDH Consulting for hosting the meeting.

Abstract:

In this session, SQL Santa (Allen) will review configurations in SQL 2008R2. These options allowed dynamic configuration across environments. Next, we look at SSIS 2012 and the new dynamic options, including Parameters, Expressions and Variables. This upgrade takes SSIS from simply confusing to downright confounding. We will simplify these options so any elf in the workshop can follow along.

SQL Integration Server developers need to make dynamic changes to properties and connections when moving packages across environments.

‘Configurations’ enabled SQL Integration Services 2008 R2 developers to make those changes, but it could be very confusing when multiple Configurations were used in combination.

SQL Integration Services 2012 maintains support for Configurations while adding a lot of power and versatility with Parameters, Environments and Variables.

References:
Professional Microsoft SQL Server 2012 Integration Services
by Brian Knight, Erik Veerman, Jessica M. Moss, Mike Davis, Chris Rock
2008R2 – Defining a Configuration Approach
http://msdn.microsoft.com/en-us/library/cc671625.aspx
2012 SSIS Parameter Basics
http://msdn.microsoft.com/en-us/library/hh213214.aspx
2012 SSIS Catalog Basics and  list of stored procedures
http://msdn.microsoft.com/en-us/library/hh479588.aspx

SSRS Data Alerts

This is definitely a good improvement from the user perspective.   It allows users (OK, they may need help) to set up subscriptions based on criteria in the report.  An example might be only sending the AdventureWorks  sales report when the sales amount was over $x or was in a certain region.  I like this feature so much, I’m starting a presentation about it.

Power View for Tabular Models

The PowerView add-in allows report authors to build interactive reports against data from PowerPivot workbooks in a PowerPivot Gallery or tabular models deployed to an SSAS 2012 instance.   For a good overview see: http://technet.microsoft.com/en-us/library/hh213579.aspx

For a tutorial:  http://technet.microsoft.com/en-us/library/hh759325.aspx

SharePoint Mode – Re-architected

This is an under the hood improvement to improve performance, management and content management.  Probably the biggest improvement here is the security integration with based on Tickets, which probably deserves an article of its own to explain.   The other note is that in Integrated Mode, all administration is done through SharePoint.

SQL Server Data Tools for for Visual Studio 2010

Yes, we finally get to escape from BIDS.  No more limited toolset, feeling like step-children while all the other developers get cool stuff for Christmas.

Excel and Word Renderer Updates

These renderers were updated to allow export to the latest versions of Excel and Word with the extensions .XLSX and.DOCX. Nice update for implementations where users want their reports in Office format.

That is it, not a lot of love for SSRS in the 2012 release.  The biggest improvement for me is to say goodbye to BIDS and jump into Visual Studio.

At the time I had Visual Studio open for a demo, so my first answer was to use Configuration Manager to store all the server addresses and then Deploy to each stored server.

Here is a screenshot of the Configuration Manager. The TargetServerURL allows you to configure different server address for each configuration.

Of course as soon as I left the session I thought of two more ways to accomplish the same thing.

First, if you have a lot of reports to synchronize between servers, download ‘Reportsync’ from the Google code project: http://code.google.com/p/reportsync/.  This tool makes it easy to synchronize a lot of content quickly across multiple servers.

Finally, my favorite that I use at most clients is ‘Reporting Services Scripter’ by Jasper Smith at SQL DBA tips.   It allows you to write scripts to move reports, data sources and all other Reporting items.  The handy thing about this tool is that you can download your content to a local foldner, edit the script and then use the scripts to push the content out to as many servers as you need to at the time.
Get it here: http://sqldbatips.com/showarticle.asp?ID=62

Well, this article is one of those “I should have said…” but I hope this helps better address the issue of synchronizing reports across multiple Microsoft SQL Server Reporting Servers.

The password was emailed to all class attendees.

Project Initiation.
During the initiation phase it is important to capture time/efficiency savings as well as what I call a ‘report value’. Each of these will be used for a final ROI or Payback Period calculation.

Cost Savings

Cost savings or user efficiency savings are the easiest to track. When you or the Business Analysts (BA’s) gather requirements be sure your interview form has these simple questions about each report:
A. How long does it take you to gather the raw data for this report?
Often the raw data has to be requested from IT or someone with access to a source system. Go to them and get details about the process.
B. How long does it take you to put the raw data into the final report format?
Once the data is gathered, someone has to total it, group data, filter data and format the report. Measure that time.
C. To whom do you send the report?
The bigger the distribution of a report or analysis, the bigger that report’s impact in terms of ‘report value’ addressed below.
D. What do they do with it?
So often clerical or department people spend hours and hours gathering and formatting reports to send up the management chain where someone else has to recompile all the reports into management summaries. Through multiple levels of management this process gets repeated. Each time data is ‘massaged’ it introduces chances for errors, typos or incorrect calculations.

This series of questions will build a ‘value chain’ for the information in the company.
For a sample report: IT spends 1 hour pulling data, an accountant spends 3 hours and forwards it to 3 managers.
Each Manager takes part of the report and spends 1 hours summarizing it for their department.
This results in a value chain of: 1 hr + 3 hr + (3*1 hr) = 7 hours for a report.
If the report is run weekly, then that is 7 hrs * 52 weeks = 364 hours per year.
You can then go to HR department and find out an average cost per employee hour to determine the savings. With salary, benefits and overhead we could use an average of $50.00 per hour and get a $16,200.00 time savings value.

Added Business Value

To evaluate the ‘report value’, there are several important questions to ask:
What is the value of the decisions being made from the reports or analysis?
What is the frequency of the decisions?
What is the improvement can be made with better access to information?
These three questions work together towards a ‘value of better decisions’. If a healthcare manager evaluates contracts which range from $1 to $20 million dollars once a year and a better decision could yield 5% better income, then we can calculate: $10m (avg.) X 1 time per year X .05 = $500,000.00 difference. If three managers are making similar decisions, then multiply the value times the number of managers. This methodology is based on an IBM whitepaper which regretfully is no longer found on their website.

After my first project was in production, I followed up on a healthcare CFO who was making just such a contract decision. He said he had a ‘gut instinct’ not to renew a certain contract, but by using the data warehouse and BI tools he was able to get a specific dollar answer about that contract performance over the past three years. He repeated ‘he was able‘, not his assistant, not accounting and not IT. He had information at his fingertips to make better informed decisions. I had to ask the value saved? “about $400,000″.

Discussion of Managerial Accounting

I know my graduate Accounting professor, Dr. Spann would chastise me for confusing ‘sunk costs’ in what should be a straightforward managerial accounting decision. In terms of accounting, ‘sunk costs’ are those costs which you would not change regardless of the decision being made.   For example, the employee salaries may be considered a ‘sunk cost’ because whether or not you do the BI project you don’t plan to fire the employee whose time was saved.   While that may be true, when I meet with CFO’s and heads of accounting departments they point out that by freeing up that .5 accountant and .25 IT staff, they can postpone hiring more staff for the department. Several managers have pointed out they would gladly free up their people for more ‘value added’ tasks.

 Deployment Phase

Once a BI project is deployed to the users, follow up by the business analysts will help validate several things:  did the project meet their needs? how much time does it save them?  how much value does it add to their analysis and decision making?  Going back to ask these questions helps confirm what was gathered in the project initiation phase.   Building a feedback loop into the project will also ensure that there were no missed requirements or ‘late arriving’ requirements.