zenvisage

Visual data exploration can be tedious, manual trial and error process Other Visual Analytic tools support one visualization at a time Finding patterns in large datasets is tough A lot of data exploration falls into a few key operations Composing visualization collections Filtering visualization collections Comparing visualizations Sorting visualizations What if we can automate this search for desired visual patterns?
展开查看详情

1.

2.

3.

4.

5.

6.

7.

8.Query execution ZQL queries are parsed and executed by the back end The backend translate ZQL queries into a query plan in the form of a directed acyclic graph. Two kinds of nodes: Collection Nodes Process Nodes

9.Process Function Questions What are some interesting t functions? What properties or patterns can these t functions express? What do different distance functions show? How do we know what distance metric is best? Our process structure can allow for user defined functions. What would be some useful functions that do not fall under the two functional primitives ZQL supports?

10.Zenvisage Interface

11.Optimization Techniques Why no improvements with parallelism?

12.ZQL Table Based Queries