MedGraph is a project aimed at applying Knowledge Graphs in the domain of biomedical papers. To achieve this goal, MedGraph leverages Named Entity Recognition (NER) technologies to identify and classify entities mentioned in the abstracts of biomedical papers. These entities, along with their connections, are then represented in a knowledge graph.
The resulting graph not only facilitates the understanding of existing relationships between the papers but also allows for interactive and dynamic exploration of the data through a dedicated web app.
In order to run the browser tool, please install the following libraries:
pip install -r requirements.txtThen, run app.py and open the development server.
⚠️ Please keep in mind that, in order to correctly display nodes information, you need to be connected to the MedGraph database, currently local hosted. ⚠️
- Mouse-wheel: Zoom.
- Left-drag: Rotate.
- Right-drag: Pan
- Node Left-click: Focus a node and show its links and details.
- Node Right-click: Opens the selected paper in a new tab (hyperlink to arxiv.org).
- Edge Left-click: Shows infos on the selected link and the nodes connected by it.
- Typing into the Search bar: Focus a node that was searched in Search bar.
This project has been realized for the Fundamentals of Data Science and Machine Learning exam in University of Salerno by: