[R] contingency tables

[David Winsemius]

On Mar 16, 2012, at 9:46 AM, mari681 wrote:

> Ok, before I definetly give up, and throw the laptop out of the  
> window, or
> fill my data.frame manually, I'll ask for some help.
> I have a data.frame named MyTable with 3 columns, that looks like  
> this:
>
>     V1              V2                       V3
> red-j    appearanceblood-n 105.032
> red-j   appearanceground-n  93.749
> red-j      appearancesea-n 102.167
> red-j      appearancesky-n   10.898
> orange-j	appearanceobject-n	109.354
> orange-j	as_adj_aspainting-n	93.248
> orange-j	colorbanknote-n	159.167
> orange-j	colorcar-n	117.985
> yellow-j	appearanceskin-n	109.527
> yellow-j	areacircle-n	87.064
> yellow-j	areainfarction-n	120.759
> yellow-j	as_adj_ascorn-n	219.739
> yellow-j	as_adj_asflax-n	122.576
> yellow-j	as_adj_asgold-n	90.814
> green-j	appearancewater-n	91.477
> green-j	architecturebuilding-n	103.582
> green-j	architecturecourse-n	103.325
> green-j	areaRio-n	106.614
> green-j	areaauditorium-n	102.505
> green-j	areacity-n	150.005
> blue-j	appearancecypress-n	145.133
> blue-j	appearancefirmament-n	148.655
> blue-j	appearanceman-n	85.731
> blue-j	appearancerange-n	90.706
> blue-j	appearancesurface-n	100.991
> blue-j	areagraph-n	92.708
> blue-j	arealibrary-n	77.135
> purple-j	appearanceleave-n	119.423
> purple-j	appearancelesion-n	134.287
> purple-j	colorViking-n	145.516
> purple-j	coloramethyst-n	175.619
> purple-j	colorbanknote-n	158.045
> purple-j	colorbottle-n	132.395
> purple-j	colorchocolate-n	141.833
>
> The elements on the first column can be: red-j, orange-j, yellow-j,  
> green-j,
> blue-j or purple-j. Elements on column 2 are contexts in which the  
> color
> appears. And in column 3 there is a kind of frequency measurement  
> between
> color and context.
>
> I am trying to build a new data.frame where the 6 colors in column 1  
> are
> placed as column names, all the unique contexts in column 2 are  
> placed as
> row names and the values on column 3 are in the correspondent cells  
> (or 0 if
> an intersection color-context is empty).
> Easy to prepare the empty data.frame, but I cant find out how to  
> fill it up
> with the frequencies.
> Ideas? Anybody has a ready script for this?

?xtabs
aRealTable <- xtabs(V3 ~ V1+V2, data=MyTable)
>


-- 

David Winsemius, MD
West Hartford, CT