Repository for the experiment ""Knowledge discovery in Workplace Health Promotion Practices"
To explore how scientific knowledge about Workplace Health Promotion is related, and potential gaps for further research.
- 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
andabstract
- 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
- 4814 titles extracted
- Cluster analysis with topic aggregations
- Visual representation of knowledge
- Research gap analysis
- Use python 3.6
- install dependencies with
pip install -r requirements.txt