martin-ouellet.blogspot.com
Martin Ouellet Notes: October 2011
http://martin-ouellet.blogspot.com/2011_10_01_archive.html
Anything but a. blog! Tuesday, October 04, 2011. A Source-to-Target self-documentation system. In the BI world, the corporate Data warehouse platform gets often overly complex quite rapidly. The number of source components increases rapidly to reach nearly thousand of entities:. Consider 10-15 different OLTP’s sources each having roughly 10-20 entities worth of interest (tables and views). Consider also the few File-based sources each accompanied with one or more feed entities. Now consider that each one.
martin-ouellet.blogspot.com
Martin Ouellet Notes: September 2011
http://martin-ouellet.blogspot.com/2011_09_01_archive.html
Anything but a. blog! Saturday, September 17, 2011. Oracle choices for Multidimensional Analysis. Here’s a quick assessment highlighting the main differences between Hyperion Essbase vs Oracle OLAP. Knowing that both products are now under the same ownership, I thought this should be archived before it gets completely outdated! Separate to the Oracle database. End-user focused, popular among business users as data access done via Excel. Aggregation management solution for SQL-based BI applications. Advan...
martin-ouellet.blogspot.com
Martin Ouellet Notes: Cassandra data modelling
http://martin-ouellet.blogspot.com/2015/01/cassandra-data-modelling.html
Anything but a. blog! Sunday, January 11, 2015. This post describes NoSQL Cassandra. Database solution with focus on data modelling aspect. Cassandra competes in the space of NoSQL Column Family. Storage engine, named under various terms like Wide Column Store. S storage in reference to the influential Google BigTable. Paper Competing candidate includes the HBase. Whose goals is to change RDBMS data storage from row-based to column based. Store: documents are usually stored using any open. Column Family ...
martin-ouellet.blogspot.com
Martin Ouellet Notes: March 2013
http://martin-ouellet.blogspot.com/2013_03_01_archive.html
Anything but a. blog! Friday, March 01, 2013. Data Vault model: Mobile Telecom example (part-3). Here's my last note concerning " Data Vault model: Mobile telecom example. Where some points and issues are discussed. Open Points and Discussion. 8594;Business Key inconsistencies. Among all OLTPs used as source, it is very unlikely that they will all share exact same Key(s) to represent same business entity! If we are lucky, keys will vary in format only, but more frequently, keys can be totally different.
martin-ouellet.blogspot.com
Martin Ouellet Notes: January 2012
http://martin-ouellet.blogspot.com/2012_01_01_archive.html
Anything but a. blog! Tuesday, January 10, 2012. Ever heard about Data Vault. I'll present here some key elements related to the data modelling aspect of this approach invented by Dan Linstedt. For those interested in learning more, you can also check out Dan's freely available educational video. The modelling technique proposed can leverage one of the advantage going to. In data warehousing (DWH) schema, i.e. provide more flexibility against data model changes in source. In the era where. In a very dist...
martin-ouellet.blogspot.com
Martin Ouellet Notes: July 2014
http://martin-ouellet.blogspot.com/2014_07_01_archive.html
Anything but a. blog! Sunday, July 27, 2014. We have been deploying a database service at work which is based on PostgreSQL. Here's some notes I gathered for the installation and basic configuration of the service. So far, we are quite impressed and satisfied by the stability of this "World's most advanced open source Database" (or maybe the quote should be "Oracle wannabe")! Global Wide Server Setting. This global setting defines amount of memory shared across all. Is definitely too low. Updating this g...
martin-ouellet.blogspot.com
Martin Ouellet Notes: September 2012
http://martin-ouellet.blogspot.com/2012_09_01_archive.html
Anything but a. blog! Monday, September 03, 2012. Understanding and Using SVD with large dataset. When confronted with large and complex dataset, very useful information can be obtained by applying some form of Matrix decomposition. One example is the Singular Value Decomposition (SVD) whose principles yielded the derivation of a number of very useful application in today's digitized world:. Http:/ cklixx.people.wm.edu/teaching/m2999-3f.pdf. Or Latent Semantic Indexing). In this post, I discuss the under...
martin-ouellet.blogspot.com
Martin Ouellet Notes: October 2013
http://martin-ouellet.blogspot.com/2013_10_01_archive.html
Anything but a. blog! Wednesday, October 30, 2013. At work we have won a contract in Jordan recently. This gives me the opportunity to make longer stay in this country and enjoy more than just the few days typically spent at the hotel. This is my first work experience in middle east, and there is a lot to learn being surrounded by a very different culture and habits than what you are used to… which is nice, I’m always keen in discovering other way of life. Shot taken from my hotel top floor. The whole re...
martin-ouellet.blogspot.com
Martin Ouellet Notes: August 2013
http://martin-ouellet.blogspot.com/2013_08_01_archive.html
Anything but a. blog! Tuesday, August 06, 2013. BI “ideal” platform. In dealing with BI projects, you are much more likely to work under the constraint of an existing platform than building one from scratch. This means having to deal with platform idiosyncrasies, sub-optimal architecture, complex and large data models, tangled loading dependencies and scheduling, confusion or mix of approaches and architecture, etc. Visibility: any layer can only depend on the immediate lower (preceding) layer. Data depe...
martin-ouellet.blogspot.com
Martin Ouellet Notes: June 2015
http://martin-ouellet.blogspot.com/2015_06_01_archive.html
Anything but a. blog! Friday, June 05, 2015. Spark data processing/analytics platform. At work we are looking to leverage a large scale data processing/analytics platform called Spark. Before doing some hands-on work, I will always do some research to help me get started and have better insights and context on a large scale platform. So this post summarises these notes on one of the hottest techno on Big-Data which officially past the-peak-of-inflated-expectation. Data mining and batched analytics. Data ...
SOCIAL ENGAGEMENT