Comments for Can a business intelligence product be used to answer analytic questions? at http://www2.nextanalytics.com , comment 1 to 9 out of 9 comments http://www2.nextanalytics.com Tue, 07 Feb 2012 06:54:57 +0100 FeedCreator 1.7.2 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-21 Hi Sandeep, Glad to hear from you! In response to your question, I invite you to register and download for free the developer version. All that you need to know is available on this site. It includes open source sample code, code and analytic workflow samples, several kinds of tutorials, and we have a free forum for you to ask questions and get answers. By offering a free developer environment, we are encouraging independents and ISVs to learn Nextanalytics and develop their own products; with your post, you're asking for what we provide. That's good to hear, it means we're not far from the mark in the eyes of the public. Gradually, over the months since I wrote that blog, we have found the biggest demand for our products is from people who deploy dashboards and business intelligence capabilities rather than ISVs themselves. - ward Fri, 24 Oct 2008 08:06:30 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-20 Hi ward, Best thing a tool can do in this competitive sphere is to be open for ISV to create multiple domain /technology specific products.Open source nextanalytics can be good for that.But important part will remain how many trained resources of this product exist is market to company to implement this.As this is farely new tool how can one educate himself about its functionality.and hence can compare this with others. Can we download its trail version and start exploring it.Also want to know is it open to customisation with langauges like java, .NET, perl or php.then it would be very good alternative.Is there enough example on website to educate oneself step by step.If u provide step by step learning path on this tool for BI develpoer and SDK based customisation part then it can really lead the pack. - sandeep sharma Fri, 24 Oct 2008 00:57:05 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-17 Thanks for participating in the discussion. Is this you? http://www.intelligententerprise.com/movabletype/blog/mmadsen.html It seems to me like you've taken a subset of what I wrote and then claimed it doesn't hold up as being difficult for other products to do. If all you're talking about is the ability to do moving average and comparisons, then you're right, those wouldn't be difficult at all for an olap server. But that wasn't my point. But, my point was how difficult it is for other products to do all those operations, together, in sequence, and iterate on it for each region. I know of no OLAP server that could support that and I cannot agree it is feasible in SQL or Excel? In response to my blog, Mosha Pasumansky of MSFT wrote a complex MDX statement and he did a very impressive job, but he only delivered part of it. When I asked others to finish his work (make it iterate region by region), nobody stepped up to the plate. As you may know, I used to be the Director of PowerPlay Client Development at Cognos from 1991 to 1997, and I was VP of Reporting and Analytic Products and VP OLAP Products at Business Objects from 2000 to 2003. None of the tools from those companies come even close to being able to do what I described. As for products from other vendors, I just don't know of any - how can I research them if I don't know about them? In your reply, why not tell us: "Product X from SAS could do this" or "Product Y from Oracle could do this"... feel free to educate us! People ask me who the competitors of nextanalytics are... honestly, I have trouble naming them. If there's a company who can do what nextanalytics does, I don't know them. SAS or Oracle? Maybe, but at a thousand times the cost probably. When you say "without needing a special schema" I am reminded that most OLAP users don't even consider the IT costs of their cubes. They aren't free! Not only is there the aquisition cost, but there is also the installation and ongoing population of the dimensional model. A very big expense! The fact that one even needs a schema is one of the things I'm calling into question. - ward's response Fri, 08 Aug 2008 14:32:24 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-16 The strawman you put up doesn't hold. It's fairly common in retail to do moving average comparisons. I've done all sorts of product by region year-to-year and 3-year moving average rankings in BI tools and I didn't need any special schema or code. It may sometimes be challenging to do several of these simultaneously in some of the tools, but you should research the competing tools before making claims that it's not possible, or that one needs to use code or write complex raw SQL. - mrk Thu, 07 Aug 2008 19:35:11 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-14 Hello Arjan and others! It took a while, but we finally have put the real code up there. http://www.nextanalytics.com/public/Examples/MDX-Challenge.html If you have any questions, please don't hesitate to ask! Best regards, Ward - ward (cto of nextanalytics) Wed, 11 Jun 2008 11:39:01 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-13 Hello Ward, Ofcourse I've been reading the blog posted by mosha as an anwser on your comment. Can you place the real code that you should have build to generate the pseudocode style? I just am curious if it is realy simpler to understaind then MDX when you ar pretty new to analytics. Best Regards, Arjan - ArjanF Tue, 10 Jun 2008 02:37:13 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-12 Thanks Ward. I posted my detailed answer to your question in my blog at http://sqlblog.com/blogs/mosha/archive/2008/06/08/mdx-answer-to-nextanalytics-challenge.aspx - Mosha Sun, 08 Jun 2008 20:06:47 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-11 Mosha, thanks for taking the time to make a comment. If you are who I think you are, you are possibly one of the world's foremost MDX experts. You of all people would of course you know how to do something in MDX, even if it was complex. But I think perhaps I didn't phrase the blog very well. I had meant it is "one" question, not separate questions. Allow me to describe the processing a bit better in a pseudocode style: // the column axis has time periods, the row axis has regions and products. for each region { // part 1 calculate the rolling six month average calcualte the growth of that count each "growth" calc to the row average keep only rows above average six or more times // part 2 remove all time periods except the first and last calculate growth remove any below 25 % // part 3 intersect part 1 and part 2 } Note that in part 1, the output of the rolling average becomes the input to the growth calculations. Now that I've been more clear, do you still think this is achievable in MDX? Is it fair of me to qualify your claim by saying "In a single MDX expression?" "someone with a normal amount of MDX education" and "not having to create intermediate data structures in a server or disk"? Because, if it takes any of that, then my blog's point still stands IMO. To do all that with nextanalytics is about just as many lines as what I just typed above. No external tables, no external cubes. And, the nextanalytics "code" is still readable a year later and easily modified. - ward Tue, 03 Jun 2008 14:20:23 +0100 http://www2.nextanalytics.com/nextanalyticsBlog/BusinessAnalytics/Can-a-business-intelligence-product-be-used-to-answer-analytic-questions.html#comment-10 > That’s an example of one analytic combined with another, iteratively and sequentially. Neither of which would be possible in MDX or SQL without the creation of a complex underlying data structure All these questions can be answered easily with MDX without need to do anything special, unless you consider OLAP cube to be "complex underlying data structure". - Mosha Tue, 03 Jun 2008 13:28:18 +0100