Community mailing list archives
Re: Odoo - BigData!by
On Sun, Oct 25, 2015 at 7:34 PM, Nhomar Hernández <firstname.lastname@example.org> wrote:
Reports?, Use cases?, Documentation? all of them?Then, if such libraries has the "generic" implementation for data mining concepts, "Which is especifically the objective of this project?"You have a very complete set of libraries here.Good initiative.But just FYI.Mostly datamining tools are already in Python itself and every implementation is a very especific case.It should be cool if you share your vision to understand the objective better.For example:To data mining we use "pandas" if things become more complicated we mix that with huge SQL + Pandas.
The first step in this world is understand what means "big data" and is more related about how to storage data, as you read in next post people is working with others DB systems, IMO the first step here is check in odoo how manage this.
The output of this work could be some new fields type like JSON, JSONB, HSTORE, ARRAYS ( PG types and all of us, love postgres).
For example if we get JSON field, all contacts from partner could be stored in one column or manage product attributes in the same way.
This could be the first step to get big amounts of data in new type of storage to be managed fast, some challenges  are highly recommended.
Obviosly is not just about data in PG, there is other data like logs from odoo could be analyzed, like errors and learn from this data using tools like elastic products: logstash, search and kibana
This last tools can used like mining too, and many others too but all of this will looking for data and how is stored.
This topic is really beautiful and of course how we can get new types from our postgres.
Regards and thanks for share.