🗓️ 01112024 2245
📎 #data_processing #database

ONLINE ANALYTICAL PROCESSING

Access Pattern​

  • Much lower volume of requests compared to oltp
  • Each query usually much more demanding (many million of records scanned in a short time)
  • Primarily used by BAs, not users

Dump​

  • Disk bandwidth is often the bottleneck

How does OLAP work? ​

  1. Collection - OLAP server collects data from multiple data sources (RDBs / Data warehouses)
  2. ETL (Extract, transform, and load) - Organising / aggregating
    • Use tools clean, aggregate, pre-calculate, and store data in an OLAP cube according to the number of dimensions specified
  3. Analyse - Use OLAP tools to
    • query
    • generate reports from the multidimensional data in the OLAP cube
TermsDescription
OLAP CubeData structure to organise / analyse large volumes of data (Optimised for quick and complex analytical queries)
Multidimensional Expressions (MDX)Queries for manipulating DBs

Types​

Multidimensional online analytical processing (MOLAP)​

  1. Creating a data cube - represents multidimensional data from a data warehouse
  2. MOLAP system stores precalculated data in the hypercube
  3. Data engineers use MOLAP because this type of OLAP technology provides fast analysis 

Relational online analytical processing (ROLAP)​

  1. (ROLAP) allows data engineers to perform multidimensional data analysis on a relational database
  2. Data engineers use SQL queries to
    • retrieve specific information based on the required dimensions
  3. Suitable for analyzing extensive and detailed data
  4. Slow query performance compared to MOLAP. 

Hybrid online analytical processing (HOLAP)​

  1. Combines MOLAP and ROLAP to provide the best of both architectures
  2. HOLAP allows data engineers to
    • quickly retrieve analytical results from a data cube
    • extract detailed information from relational databases

Data modeling in OLAP​

SUMMARY

Data modeling is the representation of data in data warehouses or OLAP databases

  • Data modeling is essential in relational online analytical processing (ROLAP) because it analyzes data straight from the relational database
  • Multidimensional data stored as a star_schema or snowflake_schema

How does OLAP compare with other data analytics methods?​

Data mining ​

INFO

Analytics technology that processes large volumes of historical data to find patterns and insights

Business analysts use data-mining tools to discover relationships within the data and make accurate predictions of future trends

EXAMPLE

  • For example, marketers could use data-mining tools to analyze user behaviors from records of every website visit

  • They might then use OLAP software to inspect those behaviors from various angles, such as duration, device, country, language, and browser type. 

References​