The last time I was in Boston, I got busted. I wanted clear my name, so I returned to the scene of the alleged crime to investigate, but first I had to attend to some business at the TDWI Spring
conference.
The flight from Seattle to Boston was smooth. Alaska always does a good job of getting me where I want to go. While on the flight, I used the time to get caught up on some reading. One article that
caught my eye was Hugh Watson’s “Three Targets for Data Warehousing” in the Business Intelligence Journal (TDWI, 4th Quarter 2006).
Watson, a TDWI Fellow and one of the discipline’s leading researchers, observed that companies use data warehousing initiatives for one of three reasons – for a point solution, for
infrastructure, or to support organizational transformation. Although this didn’t seem to be an earth-shattering revelation, Watson framed his findings is a very useful way. I am a big fan of
shared visions, and when an organization undertakes a data warehousing initiative, the person leading the charge would do well to ask all involved how they saw the data warehouse being used. The
ensuing discussion should tell that manager where such an initiative is going. It should also indicate what sorts of organizational changes might be needed. The article turned out to fit right in
with my first session on Sunday.
Sunday – Getting Serious about Managing BI
As many TDWI instructors have emphasized over the years, organizations will get the most out of a data warehouse when they manage that data warehouse as part of a BI program. Although point
solutions often rely on a single project, teams that depend upon that approach rarely end up with anything more than independent data marts and multiple products that address similar needs. Best
practices in BI include building a program under which the organization can launch coordinated projects. Up until now, classes have included discussions of such a program, but there hasn’t
been a class strictly devoted to BI program management. However, Mark Peco was teaching the new “TDWI Business Intelligence Program Management” class that promised to address that need
so I signed up for it.
Peco started off by covering the basics of a sound BI architecture. He then went into the specifics of program management. He pointed out that a program manager is better positioned to coordinate
issues such as general readiness assessments, quality assurance, usage monitoring, servicing, change control, and periodic value assessments than individual project managers. Peco also suggested
that we download TDWI survey spreadsheets to help us with some of these tasks. By the time Peco finished up, I had become a true fan of the course.
That evening, Sid Adelman moderated a night school session on “How BI Can Support ‘Sustainability’ and Improve Profits.” His basic premise centered on actions that a company
could take to minimize its environmental impact. He pointed out that what was good for the environment was good for business, i.e., sound asset management leads to less waste and lower costs.
Often, these companies can use BI techniques to identify these sorts of cost-minimizing opportunities as a well as analyzing employee commutes in order to encourage car pooling, etc. Members of the
class contributed to a lively discussion as the session drew to a close. The subject provided plenty of food for thought. I appreciated Adelman’s effort to expand our thinking about BI into
an area that will increasingly gain attention in the coming years.
Monday – Open Sesame
Steve Hoberman’s keynote, “Everything I need to Know about BI, I learned from My Two-Year Old” challenged us to keep the six interrogatives – what, why, who, where, when,
and how – in mind as we build and utilize our BI resources. These are things that our kids ask questions about all of the time. However, as adults, I’m sure that many of us frequently
gloss over some of these basic questions and make assumptions that get us in trouble. Hoberman also suggested a technique for personal development that I thought was pretty slick. He challenged us
to be present in what we are doing and expand our thinking with knowledge goals, i.e., learn 20 new things by lunch. I thought that given where I was – in the middle of a TDWI conference
– this would not be a problem.
Mark Madsen has spent a good part of his career following the open source software movement. I hadn’t given it much attention lately so I decided to catch Madsen’s updated class,
“Open Source Adoption in Data Warehousing and Business Intelligence.” He did a good job of tracing the interplay of patents, copyright law, and technology over the past 500 years.
Madsen made it obvious that business models for intellectual property constantly evolve over time as he prepared us for the business logic behind open source software. Open source vendors
essentially provide their software for free and make money from support services and product certification. In order to demonstrate how robust this market has become, Madsen invited four vendors in
for an afternoon panel discussion and demonstration of open source BI software. I was sufficiently intrigued by their presentations to make a note to myself to visit some of their booths in the
vendor area over the next couple of days.
That evening, Dave Wells led a night school session titled, “Turning BI into Institutional Intelligence: A BI Primer for Higher Education.” A good part of my life has been tied to
higher ed and I was well aware of the challenges faced by colleges and universities in making sense out of data collected from disparate sources with various lines of academic and administrative
drivers. According to Wells, there appears to be a very low BI adoption rate in academe. He indicated that there was no lack of critical questions that need to be answered. However, beyond politics
and diffuse constituencies, Wells suggested that we also needed to frame the advantages of BI in terms that educators would understand. He quoted Debra Friedman’s definition of Institutional
Intelligence (I2) as “the capacity of an institution to carry out analyses on questions of strategic import under continuously changing conditions….” Wells went on to point out
some TDWI resources including a white paper titled “Institutional Intelligence: Applying Business Intelligence Principles to Higher Education.” TDWI conferences might not offer pat
strategies to a particular challenge, but one can usually save his/her organization a lot of grief by avoiding things that don’t work and adopting techniques that others have found to help.
Tuesday – Time to Investigate
I decided to take a break from classes and do a bit of exploring around town. Besides, it was time to return to the alleged crime scene. Several years ago while on a Swan Boat cruise on the lake at
Boston Public Garden, I heard a didgeridoo and drum at the water’s edge. Drawn to their rhythm, I debarked and joined them with my wooden flute. We jammed for awhile until an officer of the
law told us that it was against the rules for us to be in the Boston Public Garden. Since I was, by that time, tired, I said adieu and went on my way. Now I was standing in front of the brass
plaque at the entrance to the gardens. Nowhere did it say that flute, didgeridoo, and drum playing were prohibited. My companions and I had been unjustly kicked out some years ago! Oh well, I
chalked it up to the “life isn’t always fair” column and walked back up Newbury Street in search of jazz.
Boston is home to the Berklee College of Music whose alums often go on to become headliner jazz musicians. Just outside Stephanie’s on Newbury, Ian and Jeremy were hard at work playing jazz
standards on trumpet and guitar. I stopped by for a chat and found out that they were juniors at Berklee. They really swung.
Back at the conference, I dropped in on the vendor show and followed up with BI open source vendors Pentaho and Jaspersoft. Both of these vendors had been in Madsen’s class the day before.
Both offer a suite of BI tools from basic reporting to ETL support. However, they have different business models. Jaspersoft offers open source and what they call Professional which builds on open
source software and adds functionality for commercial server platforms. Pentaho has used that model in the past, but is now moving toward opening up their entire suite in an open source offering.
Based on their demos, these vendors should be included on any organization’s list of possible BI vendors. I was just about to leave when I ran into some old friends at the Dataupia booth.
Dataupia is a new product on the market. It is what I would call a database support server. It provides inexpensive, self-healing, RAID 5 storage plus a highly scalable architecture with some
unique features. While database appliance vendors have relied on open source DBMS software that until now has suffered from a lack of BI-oriented performance features, Dataupia extends the
performance and capacity of the major BI database vendors such as Oracle, Microsoft, and IBM. It’s worth a look.
Wednesday – Star Schema Time
Even after years in the trenches and countless discussions about this technique or that, I never pass up a chance to hone my dimensional modeling skills. Chris Adamson’s class titled
“Dimensional Modeling: Advanced Topics” gave me an opportunity to do so. After a brief review of the basics, Adamson dove into multiple fact tables, advanced fact table design,
dimension table outriggers, mini dimensions, and scaling issues. I get cranky when I encounter folks who purport to know dimensional modeling and preach techniques when they lack experience. I
didn’t need to worry with Adamson. The guy was a genuine dimension modeling guru with a couple of books, lots of years of experience under his belt, and a clear ability to communicate his
perspectives to the class. My understanding is that TDWI is going to be offering more of these advanced sessions in the future. I would heartily recommend them.
That evening I sat in on a peer networking session on the “Pros and Cons of Independent Consulting” led by Maureen Clary. Clary has dealt with independents for a number of years, and
she shared some of her experiences with us. She and several old salts in the room suggested that newbies focus on marketing skills that make them unique, create an identity for their company, get
certified, i.e., CBIP, get help from a good accountant, look for opportunities to sub-contract to other consultancies, prepare for the ups and downs in their business, and develop nerves of steel.
I thought that some of the newbies had a good shot at making it. Others would be better off looking for positions within larger companies.
A couple of friends and I had agreed to meet at a local restaurant for dinner after night school. I Googled the place, read a review, got the address and headed out. It was a dark and stormy night,
and I was glad that I didn’t have to walk too far. However, when I got to the address, I found that the spot had been taken over by another restaurant about a year ago, and it was booked up.
You can learn a lot from the Internet. Of course, some of what you learn might be sorely out of date. There was a lesson in this somewhere. Was it trust but verify? What a way to get ramped up for
the next day.
Thursday – Talking about Search Engines and the Unstructured Data They Crawl…
The schedule showed Phillip Russom’s keynote, “Unstructured Data and Search: New Additions to the BI Technology Stack” for that morning, so, given my recent experience, I was
eager to see what he had to say. Russom started off with the results of a TDWI survey and a refinement in data classification. According to the survey, respondents said that unstructured data
– emails, text documents, spreadsheets, etc. – made up 31% of all data in their organizations; structured data – contained in database management systems –made up 47%; and
semi-structured data – XML documents plus various data exchange streams in industry standard record formats, made up 22%. Russom claimed that these last two categories were important
components of BI if BI is to facilitate a complete view of the organization. He then went on to explain how text analytics tools work and how BI search can be deployed against reports, report
metadata, non-BI content, and structured data. Although Russom’s keynote didn’t address the problem of data currency, which I think will be a challenge especially when drawing
information from unstructured sources, he did bring up a number of interesting points.
After the keynote, I headed for Seth Grimes’ “Integrated Analytics: Text and Data” for a more in-depth discussion of unstructured data in the BI space. Grimes was up to the task.
I was particularly impressed with his insights into the subject. He quickly moved us past the pat answer to content and tabular data integration through portals and challenged us with the question:
What do we want from our analytical systems? He suggested that it wasn’t data or some document. What we want are answers to our questions. Portals and search engines give us only part of the
picture. He pondered the proposition that search is a failure of design. He then moved on to the challenges of discovery and text analytics. Peppered throughout Grimes’ class were examples,
screen shots of various software applications, and questions, always questions. His session was intense, and I found myself later on looking at his notes and being engaged in the subject all over
again.
Friday – Flight Home
On the flight home, I broke out the TDWI Conference brochure for San Diego in August and started planning. It is one of my favorite conferences. Friends from all over the country show up, the
weather is usually top notch, and the sailing on San Diego Bay is not to be missed. See you there.
