https://github.com/matthieugomez/winsorize

winsorize observations based on interquartiles
https://github.com/matthieugomez/winsorize

Last synced: 3 months ago
JSON representation

winsorize observations based on interquartiles

• Host: GitHub
• URL: https://github.com/matthieugomez/winsorize
• Owner: matthieugomez
• License: mit
• Created: 2015-04-10T19:14:13.000Z (over 9 years ago)
• Default Branch: main
• Last Pushed: 2021-07-30T17:53:27.000Z (over 3 years ago)
• Last Synced: 2024-05-31T21:00:38.685Z (6 months ago)
• Language: Stata
• Homepage:
• Size: 17.6 KB
• Stars: 2
• Watchers: 4
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: license.txt

Awesome Lists containing this project

README

        
The command `winsorize` winsorizes observations based  5 times the interquartile.

- By default, outliers are defined as observatations with a distance to median higher than 5 times the interquartile. You can also use the option `p(pmin pmax)` to define outliers as the values below and above the specified percentiles. Use `0` for `pmin` or  `100` for `pmax' to avoid defining outliers in one direction

- With the option `replace`, variable is replaced by top coded one. With the option `gen`, a new variable is created

- with the option `missing`, outliers are replaced by missing values rather than top coded

- with the option `by`, outliers are defined within groups defined by the list of variables in `by`

The overall syntax is 

```

winsorize [varlist] [if] [in] [, p(pmin pmax) replace gen(varlist) trim by(varlist)]

```

```

sysuse nlsw88.dta, clear

winsorize hours, replace

winsorize hours, p(1 99) replace

winsorize hours, p(0 99) replace

winsorize hours, gen(newhours)

```

## Installation

```

net install winsorize, from("https://raw.githubusercontent.com/matthieugomez/winsorize.ado/master/")

```

If you have a version of Stata < 13, you need to install it manually

1. Click the "Download ZIP" button in the right column to download a zipfile. 

2. Extract it into a folder (e.g. ~/SOMEFOLDER)

3. Run

	```

	cap ado uninstall winsorize

	net install winsorize, from("~/SOMEFOLDER")

	```