Learning SSAS

From Hobowiki
Jump to navigation Jump to search

Bellevue College Info


  • Key metrics for business determines measures.
  • How do these factors need to be evaluated and this determines dimensions.
  • Only build hiearchies you need to improve speed.
  • Make a decision about the dimension, does it change slowly or rapidly? Slowly changing dimensions (SCD) let you maintain history, there should be few rapidly changing dimensions (RCD).

Lab 1

  • Generating Tables for Datamart


Lets you transfer data from one source of data to another. Here's a list of data sources:

  • ADO
  • Cache
  • Excel
  • File (word doc or otherwise)
  • Flatfile
  • FTP
  • HTTP
  • MSMQ
  • SSAS
  • Multiple files
  • Multiple flatfiles
  • ODBC
  • SMOServer
  • SMTP
  • SQL Mobile
  • WMI

MDX Queries

First off, using With is similar to creating variables in SQL to hold values. It can be used to create special sets that calculate measures or just put together values from the cube.

Data Mining

  • Associations - Correlations between variables in sets of data, or market basket analysis.
  • Sequence Analysis and Prediction - Patterns in data, similar to web based click through goals.
  • Naĩve Algorithm - All attributes have same importance.
  • Decision Tree - You feed it data to get info out.
  • Time Series - Accurate forecasting based on previous historical data.

You use DMX to run a singleton prediction against model to tell you a prediction model.