Hi Steve or other UCINet colleagues, I have the proper matrix now, but to count frequencies, I have to Transpose it.? That should be easy?enough.? Yet, when I apply the transpose command, the output only gives me the ##h file, not the ##d file.? Any suggestions on how to fix this? (using UCINet 6.738 32 bit). Thanks Jeff Jeffrey Broadbent
Professor Emeritus,? Department of Sociology Fellow, Institute on the Environment University of Minnesota Email: broad001@...¡°The world is much more interesting than any one discipline.¡±? ¨C Edward Tufte
On Wed, Jan 26, 2022 at 5:33 PM Jeffrey Broadbent via <broad001= [email protected]> wrote: That was exactly the problem.? I repasted the data properly and it worked Thanks so much! Jeff On Wed, Jan 26, 2022 at 4:29 PM Steve Borgatti < sborgatti@...> wrote:
That probably means you were not pasting from an excel file where the two values were in separate columns. If you like, send me the data and I can see for sure what¡¯s happening.
sborgatti@....
?
steve
?
?
?
thanks.? Your suggestions and the help text worked, and I got a matrix.? Except, the only problem is that the row and column actors are wrong.
The row actors for instance?read "actor1actor2", then actor1actor3, and so on.? It takes the pair as the actor.? Can you suggest how to fix this?
Jeffrey Broadbent
Professor Emeritus,?
Fellow, Institute on the Environment
University of Minnesota
Email:
broad001@...
¡°The world is much more interesting than any one discipline.¡±? ¨C Edward Tufte
?
?
On Wed, Jan 26, 2022 at 4:05 PM Steve Borgatti <sborgatti@...> wrote:
Correct: if the data are directed then you do not check the force symmetry box. When you press Save, the data are converted to a matrix and saved as
a ucinet system dataset. But only if, where it says ¡®data format¡¯, you have chosen edgelist1. That¡¯s how it knows it is supposed to convert two columns of node ids into a matrix.
?
steve
?
?
?
thanks for this guidance. ?two questions on this
1. the column actors choose the row actors, but not the reverse.? So the data is directed, right?? In that case, I do not check the force symmetry
box, right?
2. Once I get the two column list into the DL editor, how do I save it so that I can open it as a matrix in ucinet?? I tried saving it as a DL matrix,
but that does not seem to work.
Jeffrey Broadbent
Professor Emeritus,?
Fellow, Institute on the Environment
University of Minnesota
Email:
broad001@...
¡°The world is much more interesting than any one discipline.¡±? ¨C Edward Tufte
?
?
On Wed, Jan 26, 2022 at 1:47 PM Steve Borgatti <sborgatti@...> wrote:
Hi Jeff, if your data look like this in excel:
?
?
Then highlight the two columns (or three if valued data), and paste into ucinet¡¯s ¡®dl editor¡¯ (4th button from the left on the toolbar).
?
?
Set ¡®data format¡¯ to ¡®edgelist1¡¯. If you copied in a header row (in the example, I have ¡®from¡¯ and ¡®to¡¯ as headers), then check ¡®col headings¡¯ on the
right. That tells ucinet that the first row are not data.
?
The only other thing is to determine if your data are undirected and only unordered pairs are included in the data (in other words if there is a tie
between A and B, then the A B pair is include and the B? A pair is not included). If that¡¯s the case, check ¡®undirected (force symmetry).
?
steve
?
?
?
?
?
I have an excel dataset for a matrix.? I wish to import it into Ucinet as a matrix.? However, the data is listed in two?columns, egos (col 1) and
alters (col 2).? Under egos, for instance, there are 7 rows for actor?1, each listing a different alter actor.? Then the ego column goes on to list actor 2 a number of times with its alters, and so forth.? How do I import this data into UCINet so that it becomes
a proper matrix upon which I can run analyses?? Does it require a certain DL language at the start of the excel file (I could not find?the DL commands for that kind of data in the help file)?? What is that DL wording and placement in the excel file?
Jeffrey Broadbent
Professor Emeritus,?
Fellow, Institute on the Environment
University of Minnesota
Email:
broad001@...
¡°The world is much more interesting than any one discipline.¡±? ¨C Edward Tufte
?
?
I have a new 2-mode dataset where there is significant differences in scales in the columns (single vs 3 digit values). If I convert to 1-mode via sums of cross products that creates
very large values and in NetDraw I have to prune by the hundreds of ties. So, I normalized the original dataset (min-max, all from 0->1) but after similar conversion to 1-mode a single pruning in NetDraw breaks all ties, i.e., multiplying fractions makes smaller
fractions. I like Bonicich'72 but it also creates values of 0-1. By trial and error, it seems multiplying all initial values by a constant (1.5-2) allows reasonable pruning in NetDraw. Using values is definitely different than categorical binary data. Any
suggestions on how to proceed?
Jeffrey Broadbent Professor Emeritus Department of Sociology Fellow, Institute on the Environment University of Minnesota Email: broad001@...CV: http: Project on Comparing Climate Change Policy Networks:
|
