How to compute the distance between some key centroids and other ones, using Harversine formula?

[Michael Duarte Goncalves]

Hi Statalist Community,

I have to compute the distance between centroids for some municipalities in my sample. But it is the first time that I am doing this, and I do not know how to compute the Harversine formula.

Basically, I want to create a dataset and keep municipalities that are betweeen 50 and 200 kilometres from key municipalities that I have in my sample.

Here is a -dataex-:

* Example generated by -dataex-. For more info, type help dataex
clear
input float cmunine str47 munname double(lat lon)
1001 "Alegría-Dulantzi"       42.83981158 -2.51243731
1002 "Amurrio"                 43.05427776 -3.00007326
1003 "Aramaio"                 43.05119653 -2.56540037
1004 "Artziniega"              43.12084358 -3.12791718
1006 "Armiñón"               42.72326199 -2.87183475
1008 "Arratzua-Ubarrundia"     42.89116171 -2.63878816
1009 "Asparrena"               42.88968768 -2.31670754
1010 "Ayala/Aiara"             43.07768283 -3.04306746
1011 "Baños de Ebro/Mañueta" 42.52925602 -2.67863583
1013 "Barrundia"               42.91537538 -2.49407452
end

For example, let say that I am interested in centroid distance between "Amurrio" and the other municipalities.

• How could I compute the Harversine formula and compute the distance in stata? Is there any command that allows me to do that, please?

Thank you very much in advance for your help.

Best,

Michael

[Andrew Musau]

ssc install geodist, replace
help geodist

[Michael Duarte Goncalves]

Hi Andrew Musau,

Thank you so much for your answer. Everything works perfectly well.

Here is the code that I implemented:

    clear all

    use "./municipalityCoordinates.dta", clear

    * Define the municipalities for which to calculate distances
    local municipalities "Alcázar_de_San_Juan Alcalá_de_Guadaíra Carmona Escatrón Totana"

    * Loop over the specified municipalities

    foreach mun of local municipalities {

        * Get the coordinates of the current municipality

        local munname : subinstr local mun "_" " " , all

        qui su lat if munname == "`munname'", meanonly

        local lat1 = r(mean)

        qui su lon if munname == "`munname'", meanonly

        local lon1 = r(mean)
        
        
        * Calculate the distance using geodist

        geodist `lat1' `lon1' lat lon, generate(`mun'_km) sphere
        
        label variable `mun'_km "Centroid Distance from `munname'"
        
    }

Thank you very much for your help, again.
if my code is not right, don't hesitate to give me feedback and correct me if necessary.

Michael

[Mike Lacy]

Stata's native -search- command often (though not always) produces useful results if one simply searches with a keyword that would appear in a post. In my experience this is easier than searching the past content stream of StataList, which -search- does *not* do. In the current situation,

search haversine

reveals several potentially helpful programs, including -geodist-, a useful, efficient, and friendly package.

-search- takes seconds, and skimming its results takes minutes, so I'd encourage more people to give it a try for virtually any question.

[Michael Duarte Goncalves]

Hi Mike Lacy,

Thank you so much! I didn't know about this nice feature of search.
Generally, what I do is exactly that: I start searching for topics by keyword on StataList. Sometimes this works well, but unfortunately many other times it doesn't.

Thank you for sharing this very useful tool with me!

All the best,


Michael