Data Explorer Discussion

DataFrame Explorer

Use-case:

df = pd.load(...) | sparkCtx.read(...)

%%HTML
<urth-data-explorer ref="df"></urth-data-explorer>

OR

urth.widgets.explore(df)

What happens?

1. A query area created based on the df that allows Filter, Groups and Sorts

   • auto generated with knowledge of column names and potentially column values

   UI widget Match any or all [ columm |V] [ operator |V] [ value |V] +-

   This should be done as an independent element

2. A dataframe that will recieve contain the declarative query elements generated/managed by the element in #1

3. A visual configuration

• • type
• • other common setting
  • specific settings based on type

4. The visual showing the data

A bit more detail/questions

The will need to be able to inspect the dataframe. This may require access to the kernel. Not a purely browser compinent.

What does the query need to know to build itself?

• dataframe columns
• types of the columns
• A sample of distinct values for a column

Question: Should these be available through the api? Details of the dataframe are done lazy.

• getTypeForColumns()