search

Caleydo / lineage

Tree-based graph visualization: genealogies and general graphs 💬 @cnobre
BSD 3-Clause "New" or "Revised" License
22 stars 5 forks source link

Table: Allow for 'super categories' (for ICD9 code for example) that can then be expanded to sub categories #15

Open cnobre opened 7 years ago

cnobre commented 7 years ago
super_sub_categories
anniecherk commented 7 years ago

Two thoughts on implementing this:

  1. When a column is expanded, use a very light grey background rectangle to group them together.
  2. Expansion mechanism: a. Have a "+" glyph at the base of expandable columns (analogous to sorting number columns). Keep a "-" glyph at the base of expanded columns to collapse them back down b. Defer expansion mechanism to the panel. image

Requesting thoughts on whether the background rectangle seems constructive or distracting, and on which mechanism to use.

alexsb commented 7 years ago

Agree on the plus. For the - I'd suggest to use a horizontal label "Mental Disorders" on top of the three expanded categories. This would also be a good place for a "-".

Background sounds good - whether it should be a frame or a background - that's easier to judge once we see it.

anniecherk commented 7 years ago

A few things to consider, both in the context of our current data set & for a general data set:

new doc 2017-02-21 11 17 28_1

(These assume the categorical encoded described in #10)

1) Aggregating columns with "extent" qualifiers like "slight" vs "serious". Even if these columns use the same categories (ie "Y", "N"), aggregating them might be misleading because the categories actually representing different things.

2) Aggregating columns which might have been reverse encoded. It's possible that there are conventions for types of data the you record a positive absence (ie, "never"). Aggregating these columns can similarly be misleading because even if they use the same categories they're representing different meanings.

3) Aggregating columns that do not use the same categories. There are a few ways I can imagine handling this but they're a little awkward.

alexsb commented 7 years ago

We have aggregation of codes in the pre-processing now. Let's do this dynamically later.