[Notes] CSCI 585 discussion 5 BI

09 Mar 2020 | CSCI585, English/英文, Note

What’s the ‘big idea’?

• there is already so much data, take use of them

fractal curves (eg. Koch)?

• every node of the structure can have a same structure

ROLLUP vs CUBE?

• Generally CUBE and ROLLUP are used for reporting purposes and they do the Subtotal and Grand total. CUBE generates a result set that shows aggregates for all combinations of values in the selected columns and ROLLUP generates a result set that shows aggregates for a hierarchy of values in the selected columns.

Your own examples of star, snowflake dims?

Z curve - what other such curves? (can use to index)

• index 2d data to 1d
• hilbert curve, sierpinski curve

QMC?

• (Quasi-Monte Carlo method) use random number to index random number

Watch this clip: https://www.youtube.com/watch?v=eK9VUGFVDJk - shows how to do a spatial query on a map, then load the result as a new layer into the [same] map.

Watch this clip, for more examples of spatial queries: https://www.youtube.com/watch?v=5SuHxeg2dt0

Here is another take on what you learned…