Go to GitHub repository

Code description

Code for paper ; "Global reach, regional strength: Spatial patterns of a big science facility." DOI: 10.1002/asi.24811

Interactive visualizations.

To see the interactive visualizations, see the visualizations page

How it works

• Dissagregates and geocodes author addresses from Web of Science (WoS) publication data, using the column DOI to uniquely identify publications. Full functionality with any address with the following form:

[Authors] Address;  

• It also find addresses and other named locations in the Address column, but won't be able to calculate aggregation metrics without x in [x] y;
• Geocoding is done with the geopy package (More info below).

More functionality

• distance measurements between addresses and to a point of origin

Important info

This package uses geopy for the geocoding module. See geopy's documentation for more info.

• This example uses the GoogleV3 API, which needs an API Key. Read here.
• Make sure you include an API key here if needed.
• You can also change the geocoder here if you prefer.

Installation

Recommended in order to change API key and/or geocoded as detailed above.

git clone https://github.com/soderstromkr/geoaddress.git
cd geoaddress
pip install -e . 

Or

pip install git+https://github.com/soderstromkr/geoaddress.git

Note: You might need administration rights, adding --user to the end of the previous command should take care of that.

Usage

from geo_address import *

or see here for an example that includes visualizations.

Requirements

Uses python 3.x.x

• pandas
• numpy
• tqdm
• geopy
• re
• pickle
• sklearn
• haversine

Recommendations

• Current geocoder gave the best results in paper.
• begin_geocode() creates a checkpoint folder to keep track of progress, which can be re-started. (Make sure to delete the checkpoint file if doing a new run or project.)