Faster command than -geocode-

[Anne-Claire Jo]

Dear Stata users,

I would like to generate longitude and latitude information of each observation by using my dataset on ZIP code & full address variables.
I have around 2500000 observations in my dataset and when I use -geocode- command, it takes super long time.
(For instance, it took 6 hours to create 19 outputs)..
I had tried to find other commands that are much faster than -geocode-, but unfortunately I couldn't find it.
Would there be any commands instead of geocode ?


Thank you in advance,
AC.

[Jared Greathouse]

Is your unit of analysis the address level

[Anne-Claire Jo]

Hello Jared, thanks for your reply !
Nope, it's firm-level data

[Jared Greathouse]

Yeah but the addresses are meant to ID the firms right

[Andrew Musau]

If you have US ZIP codes, just download a file with corresponding latitudes and longitudes and merge these. Here is such a file: https://www.kaggle.com/datasets/joel...h-lat-and-long

[Anne-Claire Jo]

@Jared, yes it is ! But it's solved now, thanks ◡̈
@ Andrew, indeed, it would be much faster ! Thank you very much

[Jared Greathouse]

Do tell. How did you manage to solve this?