<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />Hey there,
Does anyone have a way to create a word density distribution for a column of textual data in JMP? Basically I'd like to create a summary table that has every word in the data file and a column showing how many times it appears.
Thanks!
Matt
Here's a JMP solution. This is the JMP discussion forum after all.
dt = New Table( "Word Table",
Add Rows( 9 ),
New Column( "Column 1",
Character,
Nominal,
Set Values(
{"The quick red fox jumped over the brown log",
"The quick red fox jumped over the brown log",
"The quick red fox jumped over the brown log",
"The quick red fox jumped over the brown log",
"Apples peaches pumpkin pie. If you're not ready holler I.",
"Apples peaches pumpkin pie. If you're not ready holler I.",
"Apples peaches pumpkin pie. If you're not ready holler I.",
"Apples peaches pumpkin pie. If you're not ready holler I.",
"Apples peaches pumpkin pie. If you're not ready holler I."}
)
)
);
dtcol = column(dt, "Column 1");
first_pass = dtcol << get values;
// Convert the column to one big string
first_string = concat items(first_pass, " ");
// Remove periods, change two spaces to one
first_string = substitute(substitute(first_string, ".", " "), " ", " ");
// Parse all values out of this big string, delimited by a space
all_list = words(first_string, " ");
// Add these values to a new table
wordsdt = New Table("Parsed Words",
New Column("listvals", Character, Nominal, Set Values(all_list)));
// Run tabulate to get a frequency count
words_tab = wordsdt << Tabulate(
Show Control Panel( 0 ),
Add Table( Row Table( Grouping Columns( :listvals ) ) )
);
use
proc means data =dsname n;
class varnm;
run;
give the varnm of the word variable name
otherwise use first.varnm and last.varnm
