CODEWITHSUNDEEP
rpcamatrix
Sergio
Sergio

Reputation: 763

Why USArrests data can produce PCA analysis in R when the state is non-numeric?

I've been trying to do a PCA analysis with R and 'prcomp'. My data (plan0) is a dataframe with a lot of NA's, so i do

plan0_sna <- na.omit(plan0)

The resulting data is here When I try to do 

m2 <- princomp(plan0_sna, cor=TRUE)
Error in cov.wt(z) : 'x' must contain finite values only

So, I need to convert to matrix

matrix0 <- data.matrix (plan0_sna)

but in the resulting data there is no name of the state

    head(matrix0)
   LOCAL PM2.5   BC         Al        Si
2      1    21  5.9 0.02278234 0.2993741
3      1    22  7.6 0.06149135 0.1828806
12     1    28 18.4 0.01614913 0.1905879
17     1    31 18.5 0.04290772 0.1603130
18     1    26  8.5 0.03344481 0.4836519
19     1    35 14.1 0.11562827 0.3842194

I'd like to do the analysis without loosing the name of the state, as with the USArrest data:

> USArrests
               Murder Assault UrbanPop Rape
Alabama          13.2     236       58 21.2
Alaska           10.0     263       48 44.5
Arizona           8.1     294       80 31.0
Arkansas          8.8     190       50 19.5
California        9.0     276       91 40.6
Colorado          7.9     204       78 38.7
Connecticut       3.3     110       77 11.1
Delaware          5.9     238       72 15.8
Florida          15.4     335       80 31.9
Georgia          17.4     211       60 25.8
Hawaii            5.3      46       83 20.2
Idaho             2.6     120       54 14.2
Illinois         10.4     249       83 24.0
Indiana           7.2     113       65 21.0
Iowa              2.2      56       57 11.3
Kansas            6.0     115       66 18.0
Kentucky          9.7     109       52 16.3
Louisiana        15.4     249       66 22.2
Maine             2.1      83       51  7.8
Maryland         11.3     300       67 27.8
Massachusetts     4.4     149       85 16.3
Michigan         12.1     255       74 35.1
Minnesota         2.7      72       66 14.9
Mississippi      16.1     259       44 17.1
Missouri          9.0     178       70 28.2
Montana           6.0     109       53 16.4
Nebraska          4.3     102       62 16.5
Nevada           12.2     252       81 46.0
New Hampshire     2.1      57       56  9.5
New Jersey        7.4     159       89 18.8
New Mexico       11.4     285       70 32.1
New York         11.1     254       86 26.1
North Carolina   13.0     337       45 16.1
North Dakota      0.8      45       44  7.3
Ohio              7.3     120       75 21.4
Oklahoma          6.6     151       68 20.0
Oregon            4.9     159       67 29.3
Pennsylvania      6.3     106       72 14.9
Rhode Island      3.4     174       87  8.3
South Carolina   14.4     279       48 22.5
South Dakota      3.8      86       45 12.8
Tennessee        13.2     188       59 26.9
Texas            12.7     201       80 25.5
Utah              3.2     120       80 22.9
Vermont           2.2      48       32 11.2
Virginia          8.5     156       63 20.7
Washington        4.0     145       73 26.2
West Virginia     5.7      81       39  9.3
Wisconsin         2.6      53       66 10.8
Wyoming           6.8     161       60 15.6

Why is that different?

Upvotes: 1

Views: 1698

Answers (1)

user2357031
user2357031

Reputation:

This should probably go to Stack Overflow, but I took the data from Google Drive anyway. This should work with you data:

plan0<-read.table("plan0.txt", header=T)
plan0_sna <- na.omit(plan0)
matrix0 <- data.matrix (plan0_sna)
matrix1<-matrix0[,2:ncol(matrix0)]
rownames(matrix1)<-plan0_sna$LOCAL
m2 <- princomp(matrix1, cor=TRUE)
biplot(m2)

When you convert the data frame (plan0) to a matrix (matrix0) factors (column LOCAL) are automatically converted to numbers. Therefore, there is column called LOCAL in matrix0, but it only contains numbers.

Furthermore, there are several rows for each level of LOCAL, so you can't put the LOCAL column as the row names in the data frame (plan0), since data frames do not allow duplicated row names. But, this can be done using a matrix!

So, you can first delete the column LOCAL from you matrix0, and rename the table to matrix1. Then you can assign the row names to matrix1. This would allow you to run the PCA with only numerical data, and get the names on the resulting biplot.

Upvotes: 2

Related Questions