Data Warehouses Presentation

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 7

Data Warehouses

A. Adam
Data Warehouses
Typically created for reporting, data analysis and data
mining
Read-optimized (vs. write optimized in a transactional
database)
A type of decision support system
Typically subject-oriented
Contain historical data that is nonvolatile in that it
never changes once the historical data is captured
Organization of Data in a DW
Typically organized into fact and dimension tables
Dimension tables
Contain descriptive data on an entity e.g. employee dimension
Typically does not contain numerical or measurable attributes
Fact tables
Contain numerical, measurable attributes
Have a grain i.e. one row per day per employee
Online Analytical Processing - OLAP
Differs from OLTP Online Transaction Processing which
is focused on maximum throughput for processing
transactions
OLAP deals with multi-dimensional data i.e. where more
than one dimension table is involved and one or more
fact tables are involved
OLAP analysis can be done in a traditional database but
specialized software known as cubing engines assist
with OLAP analysis by precomputing the aggregates at
each level of the dimension
Specialized OLAP software (or cubing engines) can also
enable very efficient drilldowns
Data Mining
Exploring data within a data warehouse to

Find patterns within a dataset


Discover relationships between different facets of a data set or
multiple data sets.
Patterns and relationships are validated by testing on subsets
of the data
Amazon
One the largest e-commerce retailers in the world
Very large data set of consumers buying behavior
Use Big Data technologies to
build recommendation engines
Monitor and track inventory
Handle complex logistics at its fulfillment centers
Data Appliances
Specialized hardware appliances that blend hardware
and software in a preconfigured and specially tuned
manner to handle large data sets
Oracle Exadata; one of the more popular database
appliances; marketed by Oracle as a Database
Machine.
Contains lots of fast storage coupled with specially
designed storage software plus traditional database
software
Very easy to set up, all complexities have already been
handled by Oracle, just plug and play

You might also like