tag:blogger.com,1999:blog-3721871707048712457.post4862633640508419466..comments2024-03-28T00:39:48.395-04:00Comments on Breaking BI: Predictive Analytics in Tableau Part 1: Simple Linear RegressionBreaking BIhttp://www.blogger.com/profile/02551920506874509998noreply@blogger.comBlogger11125tag:blogger.com,1999:blog-3721871707048712457.post-4797941995814977942016-06-28T22:07:25.189-04:002016-06-28T22:07:25.189-04:00Hi is there a way that we can save the fit method ...Hi is there a way that we can save the fit method some where (more precisely in Rserve session)and use the predict method for the calculated field ?SupreethMekahttps://www.blogger.com/profile/13750556903492222030noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-79213770278864795662016-05-30T06:50:43.978-04:002016-05-30T06:50:43.978-04:00What are the technical features available in Predi...What are the technical features available in Predictive analysis of BI????<br />Anonymoushttps://www.blogger.com/profile/07439309078611167460noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-36302373301214996642016-03-03T13:44:37.371-05:002016-03-03T13:44:37.371-05:00Hi Brad, nice post, congratulation.
witch sample d...Hi Brad, nice post, congratulation.<br />witch sample data set from a collegiate source did you pick? PleaseJurahttps://www.blogger.com/profile/13004341851405975218noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-58307375540856723332016-02-22T13:09:21.493-05:002016-02-22T13:09:21.493-05:00I am not able to run the lm function in the code. ...I am not able to run the lm function in the code. Do I need to install any package related to it in R first?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-57540719015771217422016-01-07T14:57:58.095-05:002016-01-07T14:57:58.095-05:00How can we run the new data to predict the depende...How can we run the new data to predict the dependent variable basis on already established predictive model?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-63073221512917759512015-12-16T12:56:35.608-05:002015-12-16T12:56:35.608-05:00Hi Brad,
I'm using a similar formula and getti...Hi Brad,<br />I'm using a similar formula and getting a strange error that I can't seem to sort out:<br /><br />Unexpected number of results returned by SCRIPT function. Function expected 5206 values; 2268 values returned<br /><br />Any thoughts on how to fix this? Here's my code<br /><br />SCRIPT_REAL("<br /><br />fit <- lm(.arg1 ~ .arg2 + .arg3)<br /><br />fit$fitted<br />"<br />,<br />SUM([Applications (2015)]),<br />SUM([Applications (2014)]),<br />SUM([Applications (2013)])<br /><br /><br /><br />)Anonymoushttps://www.blogger.com/profile/03089818407560531192noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-73215475745599273242015-12-12T07:27:58.867-05:002015-12-12T07:27:58.867-05:00how to get the percent of difference and how to ca...how to get the percent of difference and how to calculate it?Anonymoushttps://www.blogger.com/profile/10137868230811026682noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-15379186955047639952015-12-07T11:52:41.064-05:002015-12-07T11:52:41.064-05:00Yashwanth,
Just create a calculated field that eq...Yashwanth,<br /><br />Just create a calculated field that equals your parameter (ex. parameter "consumer spend", calculated field "consumer spend variable"). Initially, load the parameter with values from your field. The parameter won't have a dynamic range, but it will suffice for your needs. If you need to have the parameter be dynamic, there are ways to make it so.Bhttps://www.blogger.com/profile/17033343777750259689noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-8031011416489792772015-07-06T17:50:47.122-04:002015-07-06T17:50:47.122-04:00Hello Brad,
I have a requirement using R an tabl...Hello Brad,<br /> <br />I have a requirement using R an tableau, The details are as follows,<br /> <br />can we pass the arguments(.arg1,arg2....) as parameters, so that user can select the X and Y variables in tableau using the parameter control and the visualization changes accordingly.?<br /><br />Lets consider a Simple linear regression R script in tableau:<br /> <br />SCRIPT_int('<br />mydata <- data.frame(rooms <- .arg1, ok <-.arg2, af <- .arg3, rating <- .arg4, initial<-.arg5, time <- .arg6 );<br />lrmodel <- lm(time ~ as.factor(ok) + rooms + as.factor(af) + rating, data = mydata);<br />lrmodel$fitted<br />prob <- predict(lrmodel, newdata = mydata, type = "response")<br />',<br />ATTR([Number of Rooms]),<br />ATTR([OK or SUI]),<br />ATTR([AF Counted]),<br />ATTR([Rating]),<br />ATTR([Initial or Makeup]),<br />ATTR([total TIME])<br />)<br />Now the requirement is instead passing the variables directly for X and Y as arguments , pass the arguments as parameters??<br /> <br />Kindly let me know if this question makes sense, or if more information is needed?. Thanks awaiting for the reply.<br /> <br />Thanks and Regards<br />Yashwanthyyyrhttps://www.blogger.com/profile/03046756160758334640noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-59741888018016383952014-06-03T08:08:10.550-04:002014-06-03T08:08:10.550-04:00Ankit,
Of course! This is just R code so you can...Ankit,<br /><br />Of course! This is just R code so you can do anything you want. I think the function for returning the coefficients is fit$coeff. However, the coefficients are at the same granularity as the data. So, you'll have to get a little more clever about displaying them. Basically, the R integration is great for predicting values, but pretty clunky when it comes to testing the model because the functions can only return the same number of observations as you give themBreaking BIhttps://www.blogger.com/profile/02551920506874509998noreply@blogger.comtag:blogger.com,1999:blog-3721871707048712457.post-18523699649953274942014-06-03T05:57:37.391-04:002014-06-03T05:57:37.391-04:00Hi Brad,
Great post around regression modelling. ...Hi Brad,<br /><br />Great post around regression modelling. I just have a small questions. can we see the beta coefficients of all independent variables with dependent variable<br /><br />Regards,<br />Ankit Anonymousnoreply@blogger.com