Content in Tableau Community Forums

R and Logistic Regression with Categorical Variable

tableaucommunity@tableau.com — Mon, 03 Feb 2014 15:12:08 GMT

Is there a good reference that documents tableau and R interaction?

I'm trying run a logistic regression that contain a mix of variable times and it doens't seem to like categorical variables.

SCRIPT_REAL('mydata <- data.frame(win=.arg1, bi_prem=.arg2, cat1 = .arg3);

lrmodel <- glm(win ~ bi_prem + as.factor(cat1), data = mydata, family = "binomial"); lrmodel$fitted;',

[Win Rt BIN],[Veh BI Prem], [***])

Just to be clear the variable *** is either "male" or "female".

I'm aware I could do a conversion to logical or create a new field but for drag and dropping in real time I have to have a way to throw cateogries into the regression. I can run the same regression in R w/o an issue.

Does anyone have any indication as to what I need to do to avoid this error:

Rserve on Window VM Refusing Connection

tableaucommunity@tableau.com — Tue, 28 Jan 2014 23:16:55 GMT

We have Rserve on a Windows VM Server.

When I try to connect to it via Tableau it refuses the connection. I can ping the connection and get a return.

Error is "No connection could be made because the target machine actively refused it"

We've tried it with and without log in information.

Any suggestions?

Converting to a Date/Time Field

tableaucommunity@tableau.com — Wed, 22 Jan 2014 16:49:57 GMT

I'm working with frequency and severity data and the 'time' field we have stored is the year followed by a decimal representing the month divided by 12.

So they the fields take the form of:

1986.00000000000

1986.08333333333

1986.16666666667

1986.25000000000

etc.

How would I convert them to date/time field so I can create some xy plots?

The standard coversion lumps everything into 1905. Do I need to preprocess the field outside of Tableau?

Thanks!

Mike