https://github.com/rybesh/id-link-graph
Visualizing identifier use: who uses which IDs?
https://github.com/rybesh/id-link-graph
Last synced: about 1 year ago
JSON representation
Visualizing identifier use: who uses which IDs?
- Host: GitHub
- URL: https://github.com/rybesh/id-link-graph
- Owner: rybesh
- Created: 2016-03-03T21:11:31.000Z (over 10 years ago)
- Default Branch: master
- Last Pushed: 2016-03-03T21:30:52.000Z (over 10 years ago)
- Last Synced: 2025-02-14T14:51:30.257Z (over 1 year ago)
- Language: JavaScript
- Size: 8.79 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
This is an attempt at visualizing some of
[the data generated by our working group at the PHOIBOS2 workshop](https://github.com/jhpoelen/id-link-template/blob/master/NOTES.md). It
shows different patterns of identifier use for a few selected
specimens.
(To run this you need to have [Node.js](https://nodejs.org) and its package
manager [npm](https://www.npmjs.com) installed.)
To run:
1. Install the required libraries. In the `id-link-graph` directory:
```
$ npm install
```
1. Run the development server:
```
$ npm run serve
```
1. After a bit you should see the message `Server running at...` Now you can
point a web browser to [http://localhost:9966](http://localhost:9966) and
view. Click on one of the specimen identifier groups in the table to show
the visualization of that group.
The information about identifier links is taken from
[`data/links.json`](data/links.json), which was generated from the two CSV
files in the [`data`](data) directory using the
[`links2graph.py`](links2graph.py) script:
```
$ ./links2graph.py data/groups.csv data/id_links.csv > data/links.json
```
This relies on our manual groupings of identifiers, i.e. just us being
able to figure out that they refer to the same specimen. But it would
be more interesting to group the identifiers using something like
Levenshtein distance. It would also be good to visually distinguish
identifiers of web pages and identifiers of specimens (some
identifiers might be both). Then you could quickly scan visualizations
of the groups. What you would ideally see for each one is a hub with
many spokes: one identifier node in the center with a bunch of
document identifiers pointing to it. In practice you would see all
kinds of other shapes as this little experiment shows.