4.11 Big Data

Question 1

Big Data Has to be Big In (3x2)

Answer 1

• VOLUME
• • Must require more than one server to store all the data
• • Large amount of data
• VELOCITY
• • How fast is data being inputted?
• • How quickly do we need to process it?
• VARIETY
• • How much does the data vary?
• • Does the data vary largely in values?

Question 2

Fact Based Modelling (2)

Answer 2

• Allows you to deconstruct into units called facts
• Facts placed within a master dataset

Question 3

Master Dataset (definition)

Answer 3

Ever growing list of immutable, atomic facts

Question 4

Principles of Fact Based Modelling (5)

Answer 4

• Raw data stored as atomic facts
• Each fact must be identifiable
• Facts capture as single piece of info
• Facts are immutable
• Facts are eternally true due to timestamp

Question 5

Advantages of Fact Based Modelling (6)

Answer 5

• Simple - no indexing
• New items just added
• Data always true
• Historical queries are easy to run
• • Facts are immutable and time stamped
• Errors easy to correct by returning to earlier facts