thesis-tagger

This project can be used to tag chunks of documents with topics. In this case the documents are job ads. The project consists of a node.js server connected to a MongoDB database and a simple web interface that accesses the server via a JSON API.

Prerequisites

• Node.js
• MongoDB
• python
• jupyter / ipython

Data preprocessing

The jupyter notebook 'extractjobs.ipynb' includes the necessary steps to perform the preprocessing on the data. It is assumed here that the data is in the format as described / used in this notebook. In essence this chops up the job ads into chunks and saves the data into csv files. These can then be imported in MongoDB as described in the notebook.

Server setup

1. Clone the project
2. run 'npm install' in the 'server' directory to get the necessary dependencies
3. run the server via 'node server.js' and access the interface via http://localhost:8082/jobad-tagger/