Tutorial on Palladio from Miriam Posner →
Miriam Posner has written a really excellent tutorial on using Palladio. I'm taking notes on things to fix, but if you need an on-ramp to get started on Palladio, this is a great resource.
Miriam Posner has written a really excellent tutorial on using Palladio. I'm taking notes on things to fix, but if you need an on-ramp to get started on Palladio, this is a great resource.
An amusing take on the problem of data quality in self-service BI. See also, Steve Rumsby's musing, which kicked off the discussion.
Excellent data sets for mapping applications. The data comes in the form of Shapefiles, which can then be converted into GeoJSON format for use in browser-based mapping applications. A tutorial by Mike Bostock called "Let's Make a Map" walks through the process.
The most convincing argument I've seen so far for why RAM-focused databases are the future. Summary: Disk I/O can't keep up with CPU architectures based on exploding numbers of cores.
It's always worth being skeptical of benchmarks, but this one looks pretty good and Scala comes out looking surprisingly good for this subset of language performance.
This is an important document to reference for anyone converting between different BI reporting tools in the SAP ecosystem. It's also quite an interesting enumeration of the functionality you might expect to see in an OLAP interface.
Very good, important overview of the state of the current status of women and the career outlook for women in the BI and data-management industries. It seems to me we have made little progress recently. The anecdotes illustrate the many structural inequalities that still exist in the workplace.
Massive open online courses, or "MOOCs", are big and getting bigger. Coursera is a well known name in the space but EdX, a MOOC platform that is an expanding collaboration of universities started by Harvard & MIT, was new to me.
The New York Times ends their article about the rush of universities to join these platforms with a logic-defying turn:
As important as providing free access to students worldwide, Dr. Agarwal said, is edX’s goal of using the platform for research on how students learn, and better on-campus pedagogy.
So far, most MOOCs have had dropout rates exceeding 90 percent.
An overview of how one of the more interesting pieces of SAP BW functionality, the "OLAP engine", handles queries. From Thomas Zurek.
The idea of commoditization of a software solution space marking the debut of OSS options works fine, but it doesn’t mean that OSS will be a successful option for all solution domains. MDM for many enterprises requires sophisticated and well-defined software capabilities that support many areas including the data management practices, processes, and roles of people that must be handled well to achieve success with MDM initiatives.
via @jmichel_franco
A good take on the shortcomings of existing BI tools.
To understand the relationships between various metrics and their root causes requires significant domain expertise. This is something traditional BI providers have lacked, in my opinion. It also requires more sophisticated analytic techniques be applied where appropriate, some of which today are beyond the skills of typical business analysts.
This evaluation around domain expertise and sophisticated techniques is, I think, especially relevant in the badly-named "predictive analytics" area, where the skills are severely lacking and general purpose tools are trying and failing to make up the difference.
An interesting evaluation of the strange market activity on August 1st, 2012. See also the more in-depth evaluation here.
The moral of the story: Don't deploy test software into a production environment.
Interesting and insightful discussion of the role that stress can play in our daily lives. Especially relevant for its bearing on the debate around aggressive standardized testing in the US. But also very important to consider in a professional context whenever artificial high-pressure situations are introduced (think of job interviews).
Good commentary and analysis of Everyblock's bizarre instant-shutdown.
It appears that NBC has made the decision to shut down EveryBlock. EveryBlock was originally funded by a grant from the Knight Foundation, and when that grant ended it was acquired by NBC. That, somewhat controversial, move may have pushed the Knight Foundation to rethink some of its grant terms.
Given that history, it seems especially unfortunate that this community resource is now shut down with no warning, and the data that the community worked hard to create is now (apparently) gone for good. I hope NBC at least releases the community-created data for others to use in the future, but I doubt this will happen.
Announcement of a sharp-looking new editor meant to be used on OpenStreetMap data. Built using d3.js, funded by a grant from the Knight Foundation, open-sourced on Github, and licensed under the venerable WTFPL. So, pretty much guaranteed to be awesome.
The NYTimes is on the leading edge of data visualization - far ahead of the sort of quality and variety that any pre-packaged tools provide. (See also, 2012 US election coverage.)
Curt Monash on Gartner's latest Data Warehouse magic quadrant and the concept of the "logical data warehouse" (LDW) that Gartner pushes. It is an interesting, but problematic concept, as Monash points out.
On iOS as an enabling medium for people with visual and other impairments.