Difference between revisions of "Talk:Chart2/ChartTypes"
From Apache OpenOffice Wiki
Tonygalmiche (Talk | contribs) |
(Venn Diagram: Usefulness) |
||
Line 7: | Line 7: | ||
* Gantt Charts | * Gantt Charts | ||
* Negative value in a percent chart | * Negative value in a percent chart | ||
+ | |||
+ | [[User:Discoleo|Discoleo]] 19:15, 13 July 2007 (CEST) : Venn diagrams are very useful when plotting overlapping categorical data, e.g.: | ||
+ | * 800 patients take medicine A | ||
+ | * 170 take B | ||
+ | * 90 take both A+B | ||
+ | * in addition we may wish to display the mortalities in the 3 groups | ||
+ | |||
+ | This is even more useful when analysing more than 2 groups of data: | ||
+ | * e.g. the distribution of 5 diseases in a group of 5000 patients | ||
+ | * there are A, B, C, D, E and various overlappings of those diseases | ||
+ | * in principle, only important overlappings (> 1% of patients) should be shown | ||
+ | * again, we may add additional information, like mortality to the graph |
Revision as of 17:15, 13 July 2007
Bm@openoffice.org 15:50, 13 July 2007 (CEST) : I am not sure if Venn diagrams are really important. Is there really a need in those? Especially offered by a program like OOo chart? Maybe we could have some facility of showing overlapping objects in Draw. How would data look for a chart with two circles A and B overlapping "a bit"
Tonygalmiche 17:56, 13 July 2007 (CEST) : Me too (I am not sure if Venn diagrams are really important). My priorities are :
- Combined charts
- Surface Chart
- Bubble Chart/ Circles Plot
- Gantt Charts
- Negative value in a percent chart
Discoleo 19:15, 13 July 2007 (CEST) : Venn diagrams are very useful when plotting overlapping categorical data, e.g.:
- 800 patients take medicine A
- 170 take B
- 90 take both A+B
- in addition we may wish to display the mortalities in the 3 groups
This is even more useful when analysing more than 2 groups of data:
- e.g. the distribution of 5 diseases in a group of 5000 patients
- there are A, B, C, D, E and various overlappings of those diseases
- in principle, only important overlappings (> 1% of patients) should be shown
- again, we may add additional information, like mortality to the graph