Posts Tagged ‘powerpivot’

Microsoft’s UK Tech Days: SQL 2008 R2

I enjoy going to SQL Server community events, I usually find they provide a refreshing look at what other people are doing and provide inspiration and ideas of what I could be doing myself.  Vendor-run events are different so I attended Microsoft’s SQL Server 2008 R2 Tech Days event with mixed expectations, not sure if it was going to overly marketing-heavy or whether it really would be worth taking a day out of the office.

Thankfully I was in luck, Microsoft did a great job of treading the line between promotion and information and whilst the intro and the first 5-10 mins of each take were quite marketing oriented the majority of the content was realistic and provided honest demonstrations of the product.  Also throughout the talks presenters were offering to answer questions via SMS or via the Twitter hash-tag #uktechdays, this was a great touch and even though there wasn’t time to answer all of the questions it really added to the interactivity of the event.

First up was Power Pivot, as a product it looks to be immensely powerful and provides lightning fast analytical capabilities though I imagine it needs a decent amount of RAM and an up to date processor to achieve it – the most amazing part is that it’s a free add-in for Excel 2010!  Essentially PowerPivot allows you to extract up to a million records from a database and perform in-memory analysis with that set of data, including combining it with other data sets, combining it with data in your spreadsheet, performing calculations, making summaries, etc.  It’s well worth taking a look at the demos, PowerPivot is a massive leap forward in Excel’s capabilities but to me it seems like a step backwards in terms of the centralised BI ‘single version of the truth’ concept – allowing users to rip a million rows out of the Data Warehouse, mix them up with other data sources and then send them around via email or even publish them via SharePoint.  As it goes the Share Point integration was also pretty remarkable, allowing other users to use published reports not only for viewing but also as a data source on which to build new reports – pretty ground breaking stuff but I’d hate to be the guy debugging a report based on a report based on a report based on… (you get where I’m going).  Overall I’d give PowerPivot a 5* rating for innovation but it seems that Microsoft is using a common tactic from Formula One – trying to get ahead of the competition by taking a contrary strategy, but will it turn out like Jenson Button in Shanghai (he won) or like Lewis Hamilton in Australia (he didn’t)?

After a relatively dry talk on virtualisation and Hyper-V Live Migration (impressive stuff but I’ve seen it before) the next talk was about Report Builder 3 and having never been a user of Reporting Services I thought I was just going to sit through it and twiddle my thumbs – I was wrong.  Having been knocking about in the BI world for about 8 years or so I can really say that this release of Report Builder really cements Microsoft’s position in Business Intelligence.  It’s still not very slick from a usability standpoint but the visualisations they’ve added are stunning and having been a long-time user of Business Objects the talk actually did make me think “how hard would it be to switch?” – since I have a mature installation the answer is very hard but it still made me think.  The most impressive visual elements were the Spark Lines, Data Bars and Indicators but the maps were also pretty good especially given that you can use ESRI shape files.

The next talk was “Maximising your existing hardware CPU, memory and disk” by Ramesh Meyyappan, I’ve seen Ramesh before at SQLBits and he’s always very good, very detailed and straight to the point.  It was a great talk, taking place mainly in Management Studio rather than PowerPoint and if you get the chance to see one of Ramesh’s talkes in the future you should definitely go (but have a cup of coffee first).  Following Ramesh’s rollercoaster of a talk was much more relaxing run-through of Microsoft’s ‘database in the cloud’ offering SQL Azure, a product I find extremely interesting but don’t have an immediate use for though I expect in time as the feature-set converges with SQL Server I will be changing my mind.  Next up was StreamInsight, R2′s Complex Event Processing (CEP) solution for analysing large data streams (10k rows/sec+) on the fly without touching the relational engine – it looks interesting but I don’t have those sorts of requirements at the moment so I don’t have much of a reaction.  The day was rounded off by a presentation by Andrew Fryer about Master Data Services, a difficult topic to present in a jazzy way but it looks very interesting and if it will integrate with the spaghetti-junction of systems floating around in most organisations it could do a lot to help us keep our data warehouses in line with corporate naming conventions, it sounds like a lot of fuss over a little issue but if you’ve ever actually tried to solve the problem yourself in a company with more than a couple of source systems you’ll understand how hard it can be.

All in all a good day, I’ll give a shout out to the staff at Jumbucks in Shepherd’s Bush where I had breakfast and bought a bagful of Australian confectionary and to the Vegetarian Chinese buffet over the road for providing me with much needed sustenance.

SQLBits V: Keynote and Sponsor Talks

In November the fifth SQLBits conference took place in Newport, this time around the format differed slightly in that the regular ‘training day’ took place on a Thursday, there was a paid conference day on the Friday (only £99 for early takers) followed by the usual free Saturday conference.  If you’ve never been to one of these events they’re well worth attending and it’s great to be in the company of so many people who understand the sort of environment you work in and who freely offer help and advice.  More as a reminder for myself I thought I’d put together my personal highlights and little bits of information I learned whilst I was there.

The Keynote was given by Donald Farmer and showed took us on a journey through the challenges of BI from the perspective of a businessman from the 1920s who shared many of the same problems that we do in business today, it was quite an interesting angle and an amusing talk all-round.  The address culminated in a demonstration of PowerPivot, a new BI technology centred around fetching large sets of data into a cube sitting behind Excel then allowing the users to perform Excel-like calculations as well as joining the data to other tables/spreadsheets.  I believe that the tool is aimed at analyst-level users but the whole idea of putting large data sets in the hands of users (in the demo it was a million rows) and expecting them to manipulate it seems a little like a backwards step – isn’t the heavy lifting meant to be done on the server?  Don’t get me wrong, if the demo was anything to go by I’m not too worried about performance – more concerned that many users just aren’t equipped to (and don’t want to) perform this level of analysis.  I’ll watch with interest but colour me sceptical.

On both the Friday and Saturday I attended the Sponsor talks by Solid Quality Mentors, both given by Mark Whitehorn and both excellent.  From past experience the sponsor talks vary between product demos and general advice but Mark’s talks took a “let’s think about something different” approach.  Friday’s talk was about Social Data and considered what sort of social data large companies such as eBay and Facebook collect, how they can get the most out of it and in some cases why it can be considered more valuable than transactional data.  We also had a bit of a discussion around the moral implications of collecting social data and whether or not it was legal or ethical in certain circumstances to use the data for marketing purposes or perhaps sell it on.

In addition to working with Solid Quality Mentors Mark is also a lecturer at the University of Dundee where there will soon be an MSc in Business Intelligence, and Saturday’s talk was based around some academic research he had been involved in (published in Nature).  The research had involved manually collecting data from around 5,000 plant specimens originally sourced by John Stevens Henslow, a friend and lecturer of Darwin at Cambridge.  The study used modern BI techniques including data mining and even a Bing Maps mashup to demonstrate that Henslow had been studying variation before Darwin had arrived at his theory and was quite likely to have been a strong influence on Darwin’s own thoughts on the subject that made him one of the world’s most famous scientists.