An occasional series in which a review of recent posts on SmartData Collective reveals the following nuggets:
Trend analysis in a multidimensional environment is simply using time as one of the dimensions that you are slicing and dicing your measures by. If you want to extrapolate from data, albeit in a visual (and possibly non-rigorous manner) to estimate future figures, then often a simple graph will suffice (something that virtually all BI tools will provide). If you want to remove the impact of outlying values in order to establish a simple correlation, then most BI tools will let you filter, or apply bands (for example excluding large events that would otherwise skew results and mask underlying trends).
—Peter Thomas: “The dictatorship of analysts”
Segmentation: alive and well
I think segmentation studies are the most important type of research a company can engage in. If your company does only one piece of research this year, it should be customer segmentation to find out where to spend precious marketing dollars, and where not to. First, segments don’t have to be static, in fact I believe they should be dynamic. Part of the deliverable/output should be scoring algorithm which can be applied at any time, and this one customer can fall into a different segment if their variables on file change. Naturally, no segmentation has an indefinite shelf life. And thought should be given to updating segmentation every 3-5 years or when new better data sources become available, or when marketing tools change.
—Tom H.C. Anderson: “Death of Consumer Segmentation – Ridiculous!”
Don’t forget to archive
An archiving strategy is a critical component of data quality best practices. It will continually help you focus on improving and refining your data quality projects as well as thinking strategically about how you use and manage your data on a daily basis. Establish an archiving strategy at the forefront of your data quality initiatives and you start your efforts off on the right foot.
—Michele Goetz: “Archiving Strategy: Data Relevance”
Simplify, share, understand
I’m beginning to think we should all take a deep breath and maybe begin to frame up or organize discussions like this around themes our business partners understand – e.g., translating our findings for example across the five (and only five) essentials in business that matter; cash flow, profitability, velocity, growth, and customer intimacy… I guess what I’m driving at is a way to simplify our thinking and share what we are really doing in terms (lexicon) most of our customers and management can understand.
—James Parnitzke: “The Benefits of Business Intelligence”
Implementing custom BI
Embedding analytics is as much a business discipline as it is science. Historically, our analytics have been used predominantly by the government and scientific community to perform heavy science and engineering research. As business intelligence becomes increasingly important to compete in today’s marketplace, our analytics can now be found driving business decisions in industries like financial services, healthcare and manufacturing. Partners like Teradata and SAP are embedding our analytics into their software as a way to extend their current offerings. As their customers demand more custom BI solutions to fit unique data sets, our analytics provide a more affordable approach to meet that need. Customers now have an option to implement custom BI without incurring the massive overhead that you would typically find in a one-size-fits-all solution.
—Alicia McGreevey (in an Arjay Ohri post): “Interview: Visual Numerics’ Alicia McGreevey”
Cloud 1, Cloud 2
There are two different, but related, types of clouds: the first category of clouds provide computing instances on demand, while the second category of clouds provide computing capacity on demand. Both use the same underlying hardware, but the first is designed to scale out by providing additional computing instances, while the second is designed to support data- or compute-intensive applications by scaling capacity. Amazon’s EC2 and S3 services are an example of the first type of cloud. The Hadoop system is an example of the second type of cloud.
—Robert Grossman: “Learning About Cloud Analytics”
On a beach in Florida
This sounds silly for anyone who has ever been involved in the typical hapless library exercise of a digital “knowledge management” initiative. The lasting image for most of these efforts is of a black hole - everything goes in, nothing comes out. But get ready for a change of tune. The root of this is an exploding need among all players up and down the global supply chain to harness and leverage their intellectual property without giving up control or worse, having it hijacked and used against them. An example is the kind of engineering intensive knowledge stored in the heads of thousands of soon-to-retire manufacturing process guys in the chemicals industry. I talked to ExxonMobil about this and found they’re keen to solve it before the IP ends up sitting on a beach in Florida somewhere. At the exact opposite end of the spectrum is the issue of managing IP for a company like Hasbro. They have pure entertainment images that will manifest as profitable toys like Barbie, unless someone raids the files and dumps a load of cheap counterfeit knockoffs onto the market.
—Kevin O’Marah: “The Next Big Land Rush: Believe or Not, It’s Knowledge Management”