hregmonkey: Be A Happy Regression Monkey!

net install hregmonkey, replace from(https://weiweizheng.eu.org/uploads/hregmonkey/)
help hregmonkey

ssc install reghdfe, replace
ssc install estout, replace

Title

    --- hregmonkey (v 1.0.3) ---
                   Be A Happy Regression Monkey!  Automatically run multiple regressions with the reghdfe command and flexible variable lists, and export results to RTF files with the esttab command.

Syntax

        hregmonkey [if] [in], yvars(varlist) xvars(varlist) [cvars(varlist) idvar(varname) timevar(varname) absorb(absvars) groupvar(groupvar) indvar(indvar) aggregation(string) vcetype(vcetype) betadot(#) sedot(#) elsedot(#) quietly timer noestprint subfolder(string) save replace]

    options                        Description
    ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
    Regression Variables
      yvars(varlist)               List of dependent variables. Suggest using syntax "global yvarlist "y1 y2"" to specify.
      xvars(varlist)               List of independent variables. Suggest using syntax "global xvarlist "x1 x2"" to specify.
      cvars(varlist)               List of control variables. Suggest using syntax "global cvarlist "cv1 cv2"" to specify.

    Fixed Effects and Clustering
      idvar(varname)               Name of the individual identifier variable for fixed effects. Default is "id", which defined by command tsset or xtset.
      timevar(varname)             Name of the time identifier variable for fixed effects. Default is "time", which defined by command tsset or xtset.
      absorb(absvars)              Absorption variables for fixed effects. Default is "i.id i.time". Options include "i.id", "i.time", "i.id#c.time", "i.id##c.time", etc.
      groupvar(groupvar)           Categorical variable representing each group.
      indvar(indvar)               Categorical variable representing each individual.
      aggregation(string)          Method of aggregation for the individual components of the group fixed effects. Valid options are mean (default), and sum.
      vcetype(vcetype)              Specify one of the following three types of standard error in the report. Notice: it is not recommended to run clustered SEs if any of the clustering variables have too few t levels. A frequent rule of thumb is that each cluster variable must have at least 50 different categories.
      vcetype(unadjusted|ols)       Estimate conventional standard errors, valid under the assumptions of homoscedasticity and no correlation between observations even in small samples.
      vcetype(robust)               Estimate heteroscedasticity consistent standard errors (Huber/White/Sandwich estimators), which still assume independence between observations.
      vcetype(cluster clustervars)  Estimate consistent standard errors even when the observations are correlated within groups, which allows multi-way-clustering.  1. vcetype(cluster var1 var2) allows for intragroup correlation
                                     across individuals, time, country, etc. For instance, vcetype(cluster firm year) estimates SEs with firm  and year clustering ({it:{opt two-way clusteri 2. Interactions of the type
                                     vcetype(cluster var1#var2) i.e. where all observations of a given firm and year are clustered together.  vcetype(cluster firm#year) estimates SEs with one-way clustering (one-way clustering).

    Report Style
      betadot(#)                   Set the number of decimal places displayed for regression coefficients. Default is betadot(4).
      sedot(#)                     Set the number of decimal places displayed for standard errors. Default is sedot(4).
      elsedot(#)                   Set the number of decimal places displayed for other parameters. Default is elsedot(4).

    Print Regression Results
      quietly                      Quietly run multiple regressions with command reghdfe.
      timer                        Show start, elapse, and finish times by stage of computation.
      noestprint                   Hide the standard regression results table. Default is "show".

    Output Configuration
      subfolder(string)            Take effect with option "save". Specify where output RTF files will be saved. Default is folder "result" in the "cd" path.  Support multi-level folders syntax (e.g. "my_folder",
                                     "result\my_folder", both "\" and "/" are fine).
      save                         Take effect after install package estout. Support hregmonkey to export results to RTF files in "Chinese" with the esttab command.
      replace                      Take effect with option "save". Overwrite existing RTF files when exporting results repeatedly.
    ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


Description

    Package Name Meaning

        "h" means "hi", "hello", "happy", "high-dimension", or even "hate" multiple regressions.

    Core Advantages

        hregmonkey automates the process of running a series of regression analyses using the reghdfe command for each combination of dependent variables (yvars(varlist)) and independent variables {it:{opth x:vars(va , while controlling for additional variables specified in cvars(varlist).

        Fixed-effects are specified by a user-specified absorption option abosrb(absvars) or default variables idvar(varname) and timevar(varname).

        Standard errors are clustered by a user-specified clustering variable cluster(varname) or default variable idvar(varname).

        The regression results table can be automatically exported to RTF files by specifying the option subfolder(string) for further analysis or reporting.


Additional Resources

    Package Details

        For a comprehensive overview of the hregmonkey package, including its features and capabilities, please visit the following link:  https://weiweizheng.eu.org/talk/hregmonkey-be-a-happy-regression-monkey/

    Introduction Slides

        To view the introduction slides for the hregmonkey package, which provide a detailed walkthrough of its functionalities, read the PDF from the link below:  {browse
            "https://weiweizheng.eu.org/talk/hregmonkey-be-a-happy-regression-monkey/2024-hregmonkey%20Be%20A%20Happy%20Regression%20Monkey.pdf":https://weiweizheng.eu.org/talk/hregmonkey-be-a-happy-regression-monkey/2024-hregmonkey Be A Happy Regres key.pdf}

    Demo Video

        Watch the demonstration video to see the hregmonkey package in action. This video provides a step-by-step guide on how to use the package effectively:
            https://weiweizheng.eu.org/talk/hregmonkey-be-a-happy-regression-monkey/2024-hregmonkey Demo Video.mp4


Examples

    Install Package hregmonkey from Weiwei Zheng's Personal Website
        Note: More features will be added in the future, please update regularly to ensure a better user experience.

        . net install hregmonkey, replace from(https://weiweizheng.eu.org/uploads/hregmonkey/)
        . help hregmonkey

    Install Dependencies (Skip if has installed before)

        . ssc install reghdfe, replace
        . ssc install estout, replace

    Setup

        . sysuse auto, clear
        . gen long city = ceil(_n/20)
        . gen long village = ceil(_n/10)
        . gen long id = ceil(_n/5)
        . gen year = mod(_n-1, 5) + 2001
        . xtset id year
        . sort city village id year

    Data description

        . global y1 "weight length"
        . global y2 "price"
        . global x1 "mpg"
        . global x2 "trunk turn"
        . global cv1 "headroom gear_ratio"
        . global cv2 "rep78"

    Demo

        Run regressions for "y1" on "x1" with controls "cv1", default "id" and "year" fixed-effects, cluster by "id".
        . hregmonkey, y($y1) x($x1) cv($cv1)

        Quietly run above regressions, cluster by "city", but report regression tables.
        . hregmonkey, y($y1) x($x1) cv($cv1) clus(city) q

        Run regressions only for observations where "foreign==0", specify "village" and "year" fixed-effects, cluster by "city", and do not report regression tables.
        . hregmonkey if foreign==0, y($y1 $y2) x($x1) cv($cv1) i(village) t(year) clus(city) est(0)

        Run above regressions with the default subfolder ("result") for output files.
        . hregmonkey if foreign==0, y($y1 $y2) x($x1) cv($cv1) i(village) t(year) clus(city) noestprint s r

        Run above regressions with a user specific subfolder (e.g. "my_folder") or multiple-level subfolder (e.g. "result\my_folder", both "\" and "/" are fine.) for output files.
        . hregmonkey if foreign==0, y($y1 $y2) x($x1) cv($cv1) i(village) t(year) clus(city) noestprint sub(my_folder) s r

        Run regressions for "y1" and "y2" on "x1" and "x2" with controls "cv1" and "cv1", specify "village" and "year" fixed-effects, cluster by "city", but report regression tables.
        . hregmonkey, y($y1 $y2) x($x1 $x2) cv($cv1 $cv2) i(village) t(year) clus(city) sub(result\my_folder) save replace

Author

        Weiwei ZHENG (郑维伟)
        Homepage: https://weiweizheng.eu.org
        I am a Ph.D. candidate of Applied Economics at Antai College of Economics and Management, Shanghai Jiao Tong University.
        My research fields include regional economics, industrial economics, and spatial econometrics theory & application. My interests focus on talent agglomeration, knowledge spillover, peer effect, R&D manipulation, firm innovation, total fa ductivity, etc.
        I operate the 计量经济理论与应用研究室 "Econometric Theory & Application Seminar" (https://etaseminar.eu.org).
        I designed the best academic navigation webpage, "ETASeminar" (https://vividzheng.eu.org), to assist your academic research!
        Click for my ENG CV or CHN version.
        For academic cooperation, contact via E-mail: vivid_zheng@126.com or etaseminar@163.com.