Feature Request: Visualization to drag/drop simple summary table

btibert

When thinking about EDA, and more specifically teaching the concept of really diving into the data, sometimes its really nice to simply drag/drop a few attributes around to rows/columns and either create counts or summarize a single variable.  

The count/aggregation could be as simple as 1 per table, but ideal state, we could include row count and include another stat (row/column/total %, or an aggregation like average) per cell. 

Margin totals (row/column/total) would help reinforce the elements of "what are we looking at" and help students think about how data can take vary shapes, especially when thinking about aggregates and summaries.  

I am aware we can code something up using a variety of operators, but admittedly, it's nice to just drag/drop some columns  as you are exploring and hypothesizing (e.g. Tableau style "table" explorations) a dataset that you have never seen before.

MartinLiebig

Hi @btibert ,

just realizing, cant you miss use Turpo Prep's Pivot interface for it?

BR,

Martin

MartinLiebig

Hi,
did you see 'Extract Statistics'? Not the drag and drop but gives you an overview.
Best,
Martin

btibert

Thanks, I hadn't used that extension/operatore yet.  That said, if I were using the built in Deals dataset, I  might want to put Gender in the column, and Future Customer in the row, and have the intersection (crosstab) be the average age and or count of rows.  The Extract Statistics is output by column, which is nice, but also something we can mostly get from typical results ouput > stats.  Thanks for the idea though!

MartinLiebig

Hi @btibert ,

just realizing, cant you miss use Turpo Prep's Pivot interface for it?

BR,

Martin

sgenzer

yes @mschmitz my thought exactly. Try Turbo Prep.

Scott

btibert

Yep, that'll do it!  Thanks

IngoRM

Just if others want to check out Turbo Prep's Pivot quickly, here is a link to a video showing it: https://academy.rapidminer.com/courses/turbo-prep-data-pivoting