How to extract locations from strings of addresses?

Hi all,

I have more than a thousand of strings containing addresses. How can I programmatically extract the locations, like cities (e.g. Rome, London, etc) and countries (e.g. Italy, UK, etc)?

Thanks

Pietro

[Luna]

Could you please share an example data one or two of your strings and what kind of output you would like to see as a result?

[Guillaume]

If you are trying to parse addresses that could be written using any locale convention, then good luck. While there is some commonality, the format varies from country to country. See https://en.wikipedia.org/wiki/Address.

If your addresses have a fixed format, then a regular expression is probably the way to go, but we would need to know the exact format in order to give you a correct expression.

[pietro]

Thanks for you answers. The addresses are the affiliations reported in scientific articles, where there is no specific format. Here few examples:

Sustainable Industrial Systems, School of Chemical Engineering and Analytical Science, The University of Manchester, Manchester, United Kingdom; Dipartimento di Scienze Agrarie e Ambientali - Produzione, Territorio, Agroenergia, Università degli Studi di Milano, Milan, Italy'

Department of Agricultural and Environmental Sciences - Production, Landscape, Agroenergy, Università degli Studi di Milano, Via G. Celoria 2, Milano, Italy; Department of Earth and Environmental Sciences, University of Milano-Bicocca, Piazza della Scienza 1, Milano, Italy

Department of Land and Agriculture and Forestry Systems, University of Padua, Viale dell'Università 16, 35020 Legnaro (PD), Italy

Dipartimento di Agraria, University of Sassari, Sassari, Italy

The desidered output is:

1- Manchester, UK; Milan (or Milano), Italy

2- Milano, Italy; Milano, Italy

3- Legnaro, Italy

4- Sassari, Italy

My idea is to:

• remove stopwords
• search if any word can be found in a city database. In the past, I accomplished that by using yahoo query language, but it does not work anymore.

thanks

[Bob Thompson]

I would agree that regexp() is your best bet for performing the operation in MATLAB. It looks like you might be able to split things by commas, and then search for 'Univ' as that appears to be consistent starting point for all addresses, but without a consistent format you're going to have a difficult time figuring out exactly which pertains where.

[pietro]

Thanks, what a pity there is no wordbank for geographical locations or ways to query for locations.

[Bob Thompson]

There might be, I have never specifically investigated MATLAB for that, but I do not know of one, and do not have a specific way that I could think of sending MATLAB to the internet to search for one.

[Akira Agata]

How abou the following approach?

1. Split each line with ';' character into 1 or 2 segments
2. Split each segment(s) with ',' mark into several fragments
3. Extract the last two fragments

The following is my try:

str = [...
  "Sustainable Industrial Systems, School of Chemical Engineering and Analytical Science, The University of Manchester, Manchester, United Kingdom; Dipartimento di Scienze Agrarie e Ambientali - Produzione, Territorio, Agroenergia, Università degli Studi di Milano, Milan, Italy";
  "Department of Agricultural and Environmental Sciences - Production, Landscape, Agroenergy, Università degli Studi di Milano, Via G. Celoria 2, Milano, Italy; Department of Earth and Environmental Sciences, University of Milano-Bicocca, Piazza della Scienza 1, Milano, Italy";
  "Department of Land and Agriculture and Forestry Systems, University of Padua, Viale dell'Università 16, 35020 Legnaro (PD), Italy";
  "Dipartimento di Agraria, University of Sassari, Sassari, Italy"];

for kk1 = 1:numel(str)
  strSplit = strsplit(str(kk1),';');
  strOut = repelem("",1,numel(strSplit));
  for kk2 = 1:numel(strSplit)
    strTmp = strsplit(strSplit(kk2),', ');
    strOut(kk2) = strjoin(strTmp(end-1:end),', ');
  end
  disp(strjoin(strOut,'; '))
end

The output is as follows:

Manchester, United Kingdom; Milan, Italy

Milano, Italy; Milano, Italy

35020 Legnaro (PD), Italy

Sassari, Italy

[pietro]

Hi Akira,

thank you for the answer. Well, it is a good solution. I think I might adopt it.

Cheers,

Michele

Answer 1

[Stephen23]

Using one simple regular expression:

>> C = {...
  'Sustainable Industrial Systems, School of Chemical Engineering and Analytical Science, The University of Manchester, Manchester, United Kingdom; Dipartimento di Scienze Agrarie e Ambientali - Produzione, Territorio, Agroenergia, Università degli Studi di Milano, Milan, Italy';
  'Department of Agricultural and Environmental Sciences - Production, Landscape, Agroenergy, Università degli Studi di Milano, Via G. Celoria 2, Milano, Italy; Department of Earth and Environmental Sciences, University of Milano-Bicocca, Piazza della Scienza 1, Milano, Italy';
  'Department of Land and Agriculture and Forestry Systems, University of Padua, Viale dell''Università 16, 35020 Legnaro (PD), Italy';
  'Dipartimento di Agraria, University of Sassari, Sassari, Italy'};
>> M = regexp(C,'[^,]+,[^,]+(?=;|$)','match');
>> M{:}
ans =
  ' Manchester, United Kingdom'
  ' Milan, Italy'
ans =
  ' Milano, Italy'
  ' Milano, Italy'
ans =
  ' 35020 Legnaro (PD), Italy'
ans =
  ' Sassari, Italy'