What is it?
This proof-of-concept demonstrates:
1) how to identify important concepts within a text (= "named entities"), and
2) how to display context information when hovering over them
- For the user: Look up information on important things mentioned in the text without interrupting the flow of reading (for example definitions from Wikipedia as in the demo, or other swissinfo.ch content that provide more in-depth explanations about certain concepts, people, places, etc.).
- For swissinfo.ch: Less users leaving the site to look up something and forgetting to come back. New possibilities to aggregate content from outside sources other than simply linking out.
1-3 person days
- Based on live content from the swissinfo.ch website which is processed through a reverse proxy
- textblob for Named Entity Recognition (NLP)
Which things could be build with this POC if you had more time:
- Improve quality of Named Entity Recognition
- Provide less trivial context information on each entity