Does anyone have experience using a front-end tool like MicroStrategy with Hadoop or Aster? Is it as difficult as it seems: hardcoded SQL or Java embedded in the front end?
As far i know Microstrategy has DB connectors available for AsterData and Hadoop Hive. I think to implement map-reduce functionality Java coding is required.
What I was trying to figure out is what the user experience is like within the front end tool. It sounds like MicroStrategy to “Big Data” can’t really be done with self-service drag and drop columns. I know that the front end tool vendors have some things in the works, but I have not heard of any currently available products.
Using Hadoop for data analytics and MicroStrategy for presentation seems proven use case at some of the social gaming companies. I have worked in one of the project where Hadoop & MicroStratagy are used thought my work was mostly on Hadoop side. Due to some of the limitations with MicroStratagy we moved most of the aggregations and calculations to Hadoop.
Tableau with Hive works well!
. Does Tableau natively understand Hadoop? I.e. can you drag and drop columns to get answers? Or are you forced to hard code Java into Tableau?
common scenario will be to pull the data out from hadoop to some kind of relational analytics store (pentaho and infobright will be a good choice) and then using BI tool like microstrategy. if from some reason you would like to do an analaysis directly on hadoop checkout Karmasphere.
if u need help with you data challanges contact us at www.zaponet.com
]The way I used it, Tableau has a Cloudera Hive adapter which lets you treat Hive as yet another datasource.
You connect to this Hive ‘Table’ or Hive Query from Tableau – just the way you would with any other Relational Tableau and then start visualizing.
No I did not have to do any custom coding Java to do this 🙂
We are using Cognos in conjunction with Aster. We have users doing drag and drop with some SQL-MR functions.
In Cognos you have to set it to native SQL mode to use the Map Reduce queries. Cognos does not really have any robust variable support ( at least that we have found ) in many cases you end up having to filter Map Reduce results with wrapping SQL unless you do some hand-stands and cart-wheels.
I built Tableau’s connector to Hive, and I’m pleased to report that it offers drag / drop visualization capabilities just as we do for any other database. To manage latency, Tableau’s Data Engine is a great memory-mapped accelerator. Topics involving configuration, administration and performance tuning can be found on our Knowledge Base: http://www.tableausoftware.com/search/support/hadoop
I’m happy to answer any questions here as well.
the problem with Big data is that it is difficult to ascertain what data processing should reside in the analytic engine or the visualization tool. Tableau, Microstrategy and other BI tools can either use Hadoop’s processing power or use in-built mostly in memory data processing for complex visualizations. So an effective BI tool should have the intelligence to decide when to push the processing to Hadoop/Hive and when to use in-built engine.
Most of the BI tools support either JDBC or ODBC connectivity which allows using Hive for both drag-drop designing and executing analytics over Hadoop.
AFAIK, Intellicus- a BI tool supports Hive based report designing and we are working on a ODBC driver which would allow Tableau and later microstrategy and other BI tools to drag-design reports over Hive.
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!