As many know by now Flex has released some OLAP components. The visual components are very similar to the AdvancedDataGrid and display data in a somewhat limiting dual-axis grid. Limiting in that you will need to customized the interface to support OLAP features such as drill-down, drill-up, grouping etc. Also the components are all in-memory so you need to create flattened data sources as dataprovider sources or roll your own implementations of the IOLAP interface elements (Cube, Dimension, Query etc). We have deviated from flattening out any datasources and have actually started creating a OLAP-Flex communications library. The library run across HTTP (SOAP) using XML/A. So far the initial results are pretty fantastic. We are just testing with a Pentaho, Mondrian OLAP stack but it should not be too hard to create data sinks for Analysis Services or other XML/A implementations. Additionally we are modifying a lot of our Flex analytics tools to ingest the OLAP library and use it as a data source. The first components being released are some Toad/SQL Query browser like tools for use in executing MDX queries and browsing OLAP repositories.
Here is the intro text for the Flex 3 OLAP components:
About OLAP data grids
When working with large amounts of data, you can quickly get overwhelmed with the scope and size of the data. For example, you collect sales information for different products, in different regions, and for different customers in a typical two-dimensional spreadsheet. That spreadsheet could easily contain hundreds of rows and tens or even hundreds of columns. Extracting useful information for such a large data collection can be difficult, and trying to identify trends or other patterns in the data can be even harder.
Data visualization is a technique for examining large amounts of data in a compact format. One type of data visualization technique is to use a chart, such as a bar, column, or pie chart. Adobe® Flex® supports many types of charts. For more information, see Introduction to Charts.
Another data visualization technique is to aggregate the data in a compact format, such as in an OLAP (online analytical processing) data grid. An OLAP data grid is similar to a pivot table in Microsoft Excel. An OLAP data grid displays data aggregations in a two-dimensional grid of rows and columns, like a spreadsheet, but the data is condensed based on your aggregation settings.
Note: While the Flex OLAP data grid is similar to a pivot table, it provides a much greater set of features for aggregating data.
Being big Tuft fans we have been anxious to use sparklines in our Flex projects. After some research we found precedent in Lucas Pereira’s blog and some other miscellaneous messaging regarding sparklines in flex but overall I was pretty surprised at the lack of examples and discussion. This information void left me with the thought that maybe the Flex community is not as active in data visualization (maybe there is content with the Flex Charting components) as I had previously imagined.
Regardless of the cause for the lack of Flex sparkline examples, we needed them so lets get into what we did:
After JavaOne 2008 I started reading up about Sun’s Flex and SilverLight competitor, JavaFX. I have only scratched the surface but the brightest items thus far:
- Java-Friendly development: Using java syntax to build out a JavaFX application is a nice feature. It keeps the environment friendly to java programmers and side-steps some developers thoughts that programming in xml is beneath them.
- Database connectivity: Looking over this example for using JavaFX with a database the database connectivity features are intuitive and commensurate with traditional java-database code. Additionally, it appears as though you can use traditional jdbc protocols, no specific protocol like AMF is needed to transfer data to the presentation layer..
- Deployment: The most impressive feature, especially from a delivery stand-point, is how a user can drag the javaFX application off the web-page and onto the browser and keep the functionality in tact, even when the browser is closed. Love this. I have been trying to figure out a similar workflow with Flex and although I was successful the process still required user-intervention and an installation step. According to the demo (Beyond the Browser) the migration from browser to desktop is seamless as opposed with Flex where some tags require modification to convert AIR-Flex and vice-versa.
All in all javaFX looks like a pretty fantastic tool and it is something to keep an eye on. We will start some initial testing in the lab this quarter and blog our experience.
I was bashing my head against the monitor today trying to figure out why my actionscript created summation rows were not appearing in my Flex 3 AdvancedDataGrid component. After several hours I figured out it was, gasp, a string mismatch. I am fairly confident that my problem will appear for others so let me explain more after the Read More link.
Google is releasing Friend Connect: Easily insert social features to make “any app, any site, any friends” a reality. Great news for the tens of thousands web sites that want to become socially enabled. But I think one area where this is really interesting is in our world, where collaboration and discussions of data, conclusions and visuals are very important.
With Friend Connect or a similar tool, it may be possible to quickly empower a workforce or team to review a specific data set and work towards a mutual goal or discuss the validity of the data and keep this knowledge within the context of the data, which is where it belongs.
Of late, the most frequently asked question we have been receiving about QlikView is how to import Excel data. So I put together a quick How-To. The process, like most QlikView functions, is very simple but the starting point could be better labeled. Once QlikView is able to read your data its wizard makes it possible to quickly format, re-shape and modify it to suit your analysis needs. I would venture to say that the process is easier than importing data into Excel.
Hit Read On for full instructions
Sadly we had some hosting issues so we had to start a new. We will migrate popular blogs over in the future but until then we get to start fresh…
Although it is internal only, Amazon’s CEO is excitedly communicating their scalable, highly R&D based Dyanmo system. Judging by the paper (see link below) they really stuffed the envelope by leveraging a broad range of academic research topics. Combining all this into a production-quality, high-performance distributable persistence platform is certainly an achievement.
read more
And this improved zLinks idea seems to be a good start. zLinks enables people to annotate the links on their blogs (and other web pages I imagine) to describe the links. Once the links are described one is able to start
traversing related items ala Semantic search. One of the foundational ideas behind zLink is that a link is not just a link. To me it is trying to describe the link as a verb or action which is probably more accurate these
days. e.g. A link to a book on Amazon is not just a link but it could be a “buy book” action. Zitgist describes 3 purposes of a link.
The idea is cool and the interface is clean and very usable.
more after the jump
09 Jul 08 |