SPSS continue to develop as a solutions provider

Data Mining is now becoming very much part of mainstream
business activity. As it passes from speciality to universal usage
so it becomes imperative that the technology is hidden, and its use
becomes easier—this is what SPSS are making great strides
with. At the same time this does not mean that development is
slowing down at the level of the technology in favour of packaging,
indeed each new release can be seen to be a bigger increment than
the last.

The biggest single change that occurred inside SPSS was when
they acquired Data Distilleries. Although Data Distilleries were a
tools vendor they also had a very sizable consulting practise that
taught them a lot about what is actually required by business
people rather than the analysts who traditionally have used data
mining tools. In particular Data Distilleries has given SPSS great
insight about how to present things to a business audience. What
SPSS are now engaged in is taking all of their clever stuff that
goes on inside the data-mining engine and hiding it from the view
of the business user. The front end is now able to support both the
analyst who wants to control the model, and the business user who
wants to use the model, in very complementary fashions so that they
can work together more seamlessly.

At the same time SPSS are addressing the need to recognise that
what is required to support a “Predictive Enterprise”,
rather than the “guess and go” style of management seen
in so many companies, is an enterprise infrastructure for
predictive analysis that manages the processes of managing data and
models from cradle to grave in an integrated whole. This is a
Service Orientated Architecture that binds everything together and
allows the analytical tasks from inception to delivery to be
scheduled and run as automated processes without the necessity for
manual intervention. As all of these services become available
within an overall Enterprise Service Architecture, so events can be
monitored and reacted to closer and closer to real-time, enabling
valuable customer interaction to be fined tuned.

Increasingly, it is not just numeric data that needs to be
managed and analysed, we also require the vast amounts of textual
data captured within our enterprise processes to be analysed. SPSS
are able to apply the same rigour to text mining as they do to
handling data. Obviously context is a lot more of an issue when
developing understanding from text than it is when looking at
numbers that, for those of us not trained as accountants, tend to
only ever have one meaning. The process of extracting meaning is
split into two, having extracted the text to be examined into a
suitable format it is analysed firstly to discover concepts, and
then secondly those concepts are scored. For SPSS, text mining is
very important in the development of a true 360 degree view of
customers because SPSS are overwhelmingly the most significant
players in the Market Research market so survey data on attitudes
held textually is an abundant source of insight to a very large
number of SPSS users.

All of these changes are vital for SPSS, because it is vital to
transform the way that data mining is used in business.
Traditionally it has been a highly crafted laborious process
employing skilled staff who only had a limited bandwidth so the
numbers of models deployed by most enterprises could be counted in
tens, verging on hundreds. With the broader adoption of data
mining, and the ever-increasing velocity of business life that
model is unsustainable. Further, with Microsoft entering the market
with SQL Server in-database data mining being made available via
add-ins to Excel in Office 07, the whole world of data mining is
going to be changed forever. I believe that SPSS are ready to face
that challenge with a solid offering.