Project

Gregor Thomson - University of Glasgow

This is a collaborative project with Amazon Development Centre to investigating whether auto-generated tags can accurately represent jobs to aid job recommendations. Job Tagger is a prototype system used to suggest skill for a given job description. Input a job description and Job Tagger will return five suggested skills needed for that job.

JobTagger - Web application for the prototype system

SkillVectorGenerator - Creates a Word2Vec word embedding vector for each skill

WikipediaPopularity - Finds the average daily views for each skills wikipedia page using Wikipedia PageViews API

WikiScrape - scrapes Google for Wikipedia pages for each skill

WordSkillSimilarity - Finds words similar for each skill using Word2Vec word embeddings

=======

Running the Programs

JobTagger

An example of running from the command line is as follows:

$ mongod

In other terminal:

$ npm install
$ node index.js

Goto http://localhost:8080/ and it will take you to the index page of the website.

SkillVectorGenerator

Run the following in the command line:

$ npm install
$ node skillVector.js

WikipediaPopularity

Run the following in the command line:

$ npm install
$ node popularWiki.js

WikiScrape

Run the following in the command line:

$ python scrapeWiki.py

WordSkillSimilarity

Run the following in the command line:

$ npm install
$ node word2vec.js