Talend Open Studio Data Integration

Talend Data Integration and Management

www.robertomarchetto.com

Data Integration

Data Integration involves combining data residing in differente sources and providing the

user with a unified view of the data

Data Management combines different disciplines to manage data as a valuable resource

http://www.robertomarchetto.com/


Talend

● Talend is a company focused on Data Integration and Data Management solutions

● Talend is a „Cool Vendor“ for Gartner (2010)● Present in more than 12 locations around the

World● Fast growing company



Talend Open Studio



Talend Open Studio

● Open Source, professional tool● Draw procedures linking components, each

component performs an operation● DB vendor-specific optimized components● Produces fully editable Java (or Perl) code● Deployment with small and fast compiled Java

or as Web Service● Eclipse based IDE, excellent flexibility● BI Platform indipendent, DB Vendor indipendent



Automatic code generation, diffent deployment



Extracion Transformation Loading

● ETL is a common process in Data Integration● Extract, reading data from different datasources

(database, flat files, spreadsheet files, web services, etc)

● Transfom, converting data in a form so that it can be placed in another container (database, web services, files, etc). Cleaning, computations and verifications are also performed

● Load, write the data in the target format



Tutorial, Source data



Tutorial, Destination data (Datawarehouse)



Tutorial, Metadata

● Talend requires a preliminary definition of the metadata

● Often a strong metadata definition means, as in programming languages, fast, robust and maintenable applications

● ..demo..



Tutorial, Talend jobs basics

● Place components on the designer● Link components to build a transformation● Main type of link: Rows flow● Schema metadata is propagated and must be

coherent● ..demo..



Tutorial, users_dimension



Test the job



Tutorial, accounts_dimension



Tutorial, dates_dimension



Tutorial, write a Java library



Tutorial, opportunities_fact



Tutorial, define a root job



Deploy and run



Extensibility, comunity plugins

● Many official components

● Components for every task released by the comunity

● Geospatial components, log analysis, Google analytics, data encryption, etc



Scheduler



And now.. reports, dashboards, OLAP, Geoanalysis, KPIs..



Do you trust your data?



What about data quality?

● Customer A is present 5 times with different names

● Null values can vary statistical indexes like mean calculation

● Duplicated records● Blank values● Some records can contain errors (es -1 field

values)● Some records can be garbage



Talend Open Profiler



What abount data storage size?

● Some fields can be oversized for the data they contain

● Sometimes fields are related and can be calculated

● Some keys or values are never used● When data grow garbage grow● Data storage is not free (disks, electricity,

backups, DB licenses)



Data is „the black gold“ that can produce knowledge

● Data is a resource, you can extract knowledge● A lot of Data produces concise informations● Data storage is not free and a lot of data can

make system not fast● Data cleansing is a central process in statistical

analysis and Data Mining



Talend Master Data Management


Talend Open Studio Data Integration

Technology