Tableau dashboard text Link

A trick learned from Tableau Friction:

Replace text node in .twb file with new link node.
Example:
 <run auto-url='true' fontcolor='#d8504c' hyperlink='tabdoc:load-url url=&amp;quot;http://www.google.com&amp;quot;'>This is a link to Google</run>


Source:
Dashboard Text Links - Adding and Customizing, no Worksheets Needed 

Test change detect and breakout detect togeter

Two awesome posts in the first week of 2015.

1. Change Point Detection in Time Series with R and Tableau
2. Breakout detection in the wild

Put them together:

Updates

• A new R package for detecting unusual time series
• Time Series Anomaly Detection Algorithms

SQL script to profile data in table

Script below can be useful to produce a summary of columns in a table.
When we get a new data sources without sufficient document, this script can help us to list the number of NULLs and distinct values, as well as to list the top 50 unique values of a column.

Please download the SQL script: ProfileTable.sql

Use stored procedure in Tableau to optimize consultant practice

The typical workflow of consulting service can be labor-intensive and error-prone. We need to copy data among applications, set up calc sheet in Excel, snip Tableau screenshot, and finally to make the slides reflect the correct information scattered in modeling tools, database, and spread sheets.
What if we simplify the workflow by removing some applications, say, Excel.  
So, here is a proposal to move the calculation in Excel into SQL database with the feature of stored procedure support in Tableau 8.1 and later. 





What if we go even further, say, replace PowerPoint slides with Tableau Story Points.

Count frequency of factor column

Name: frqTable
Scenario: Want exchange data to Excel, or do the long tail fit

Code:
frqTable <- function(aFactorCol, aColName, aIsDecs="TRUE"){
  ret <- as.data.frame(table(aFactorCol))
  colnames(ret) <- c(aColName,'frq')
  ret <- ret[order(ret$frq,decreasing=aIsDecs),]
  rownames(ret) <- NULL
  return (ret);
}

Refactor: Drop unused level of a factor variable

Name: refactor
Scenario: After subset a data frame, the factor columns may include 0 frequncy levels
Related:
1. drop = TRUE, e.g. problem.factor <- problem.factor[, drop = TRUE]

2. drop.levels() function in gdata package
References:
http://r.789695.n4.nabble.com/Refactor-all-factors-in-a-data-frame-td826749.html
http://www.r-bloggers.com/data-types-part-3-factors/
http://rwiki.sciviews.org/doku.php?id=tips:data-manip:drop_unused_levels

Code:
refactor <- function (aDf){
  cat <- sapply(aDf, is.factor);
  aDf[cat] <- lapply(aDf[cat], factor);

}

Methods to save data frame in file in R

When you try to "save" your data set in a data frame object in R, you have several options:

Method	Pro	Con	Funcitons
Image the object in binary format	Fast, can keep object name and other environment information	R specific	save(df, file= "filename") rm(df) load("filename", .GlobalEnv)
Save in coded text	Full information, e.g. data mode	Size is big, can not exchange with other software	dump(c("df"),"filename") newDf = source(“filename")$value
Export  to plain text	Human readable, and software exchangeable	May need to recast R types when read in	write.table() read.table()
Export to other format	Software specific	Software specific	Write.X() Read.X(), where X can be spss, sas,csv, excel