marcelotournier / knowledge_discovery_WHP

Repository for the experiment "Knowledge Discovery in Workplace Health Promotion"

knowledge_discovery_WHP

Repository for the experiment ""Knowledge discovery in Workplace Health Promotion Practices"

Experiment outline

Objective

To explore how scientific knowledge about Workplace Health Promotion is related, and potential gaps for further research.

Methods

• Gather all article summaries from JOEM in Pubmed (adv search by Journal w. journal title)
• Parse XML data into a table with columns: article_id,article_title and abstract
• Apply NLP process: Tokenize and remove stopwords, apply Bag of Words and TF-IDF
• Run a clustering algorithm to "label" classes
• Analyze cluster results to manually add a class descriptor
• Apply t-sne to see visual representation of knowledge
• Comment t-sne results

Preliminary Results

• 4814 titles extracted

Expected results

• Cluster analysis with topic aggregations
• Visual representation of knowledge
• Research gap analysis

Project building instructions

• Use python 3.6
• install dependencies with pip install -r requirements.txt