SAS professionals' 2010

How many analytic consultants can you fit in a family hatchback?  On one of the hottest days in June, only 3.  Anything else would have just been uncomfortable…

SAS DucksAnd so it was that Tom McGuinness, Ian Smith and Duncan Court of Atom Insight travelled, in relative comfort, and with a Smörgåsbord of road snacks at hand, to the 2010 SAS Professionals Conference in Marlow.

The SAS Professionals Conference is run by, and for, SAS users.  Consequently, sales talk is minimised and interesting and interested SAS users are thrown into the same place with lots of coffee, hotdogs, beer, ducks and footballs (apparently England were briefly playing in some kind of football tournament this summer).

It all started several weeks earlier, when Tom was extended the great honour of being invited to take a slot at the conference and share his 10 ten pence’s worth on web analytics.  So it came to be that, PowerPoint slides and un-camera-friendly striped shirt at the ready, the Atom Insight boys were ready to educate the masses.

Clearly there were a great number of talks, all with useful insights, but we’ll come to those.  Tom’s “Getting the value out of your web data” talk was immediately after the 2 keynote speakers.

Getting the value out of your web data

Abstract: Tom McGuinness, Analytics Director at Atom Insight, discusses the true value of web data available in SAS CXA. By walking through real world examples that Tom has encountered in his time as a consultant in the field, he demonstrates that it takes more than the traditional web analytics views to discover the true value of your data. Examples cover marketing, operations and pricing/risk.

Top 3 take-aways:

  1. No need to create a cottage industry out of web analytics, it should be reported and analysed like any other data.
  2. More data = better.  Capture it all.
  3. Integration is key.

Here is an abridged version of the resultant video of the talk,

a marketing case study,

a form analysis case study,

a fraud detection case study,

Between us we managed to catch a view of a lot of the other talks on offer (although, alas, not Tim Harford “The Undercover Economist” on day 2).

A summarised run-down of these follows…

Forecasting Time Series with Daily and Weekly Cycles: Applications and Methods

Abstract: Professor James Taylor, Lecturer at Saïd Business School, Oxford, uses SAS software to work through Time Series Forecasting techniques.

Summary:  Phew – sounds dull, but it wasn’t.  Dr Taylor was an extremely effusive and passionate speaker, not to mention face-meltingly intelligent.

Top 3 take-aways:

  1. Experiment, play, enjoy!
  2. Keep it simple.
  3. Check whether the results make sense in “the real world”.

Upping your game with SAS9.2 and SAS Enterprise Guide 4.2

Abstract: Chris Hemedinger (SAS R&D and author of one of our favourite SAS blogs) shows off a lot of the new SAS releases’ abilities.

Summary: Although Chris works for SAS, this wasn’t a sales pitch.  Chris just loves to play, and enjoys showing off his findings (luckily for the rest of us, these often wind up on his blog – I would be stuck without his post on importing SPSS data into SAS…).

Top 3 take-aways:

  1. Regex is extremely powerful, and very cool.
  2. Analysing Twitter and Facebook data in SAS Enterprise Guide can be fun, but has extremely valuable applications to a marketing department.
  3. Enterprise Guide is a great tool for making processes transparent and easy to understand.

Wind power risk modelling

Abstract: Many countries have ambitious targets for renewable energy. One of the fastest growing sources of renewable energy is wind power. In contrast to more traditional sources of energy, wind power fluctuates greatly, which has significant cost implications for its integration into the electricity system. These costs can be reduced if the uncertainty in wind power can be anticipated ahead of time. This presentation will consider how to produce forecasts of the probability distribution of wind power. A procedure will be described that converts a forecast of the probability distribution of wind speed into a distribution for wind power. Ideas will be illustrated using data from a set of Greek wind farms.

Summary:  More from Dr Taylor, with an applied example of the techniques discussed in his keynote.

Recommendation engines

Abstract: Colin Gray of SAS shows How to use SAS to identify consumer preferences via a recommendation engine and a recent case study.

Summary:  A good demonstration of 80/20 – a simple recommendation engine is easy to build and implement.

Top 3 take-aways:

  1. Keep it simple.
  2. Using simple rankings of products bought vs. products browsed will provide a simple but effective “people that liked this, also liked…” list.
  3. Complex statistical models will make your recommendations far more targeted.

Data Visualisation techniques using SAS

Abstract: Christopher Redpath, SAS, A session focusing on using the SAS tools in conjunction with Data Visualisation techniques to help you to discern the hidden significance within your data. SAS technology helps speed the process of information handling, but it is up to the report creators to effectively communicate its meaning to the decision makers. Through the correct graphical representations of data, numbers can be brought to life to tell their stories and aid in making the right business decisions.

Summary:  Some good common sense stuff on presenting data.  Although a SAS guy, Christopher wasn’t afraid to dismiss a lot of SAS graphical output as over complicated.

Top 3 take-aways:

  1. 3D graphs suck.
  2. Think of your data-to-pixel ratio – less is more.
  3. Annotate!

Building relationships with your customers

Abstract: Crispin Westhead / Richard Hogben, Which? Ltd / Amadeus.  As Which? builds and develops an integrated CRM approach to understanding and managing the relationships with its customers, the organisation requires a SAS solution that provides a single view of their customer throughout the journey. Which?’s Crispin Westhead describes the challenges, and Richard Hogben discusses how SAS Alliance Partner Amadeus Software have helped Which? to navigate the route so far.

Summary:  A double team from Which? and Amadeus talk about working with CRM systems, and delivering the project, in equal measure.

Top 3 take-aways:

  1. Relevance and personalisation is key in CRM.
  2. Single view of the customer is critical.
  3. Strong project management is a must.

And so, as the sun started to set over Buckinghamshire, we set off homeward bound with our heads’ swimming with ideas and tummys swimming with bottled lager (or Coke in the case of our designated driver) and Satay skewers.

Will we be back next year?  You bet!