Text queries on a BIM model

Описание: A Linked Building Data #KnowledgeGraph can capture all sorts of information about a building, and with #SPARQL you can ask that graph some incredibly complex questions. But few people "speak SPARQL", and that makes it hard to investigate this rich dataset. So I decided to make an experiment with natural language text searches to SPARQL and here I present to you my first test results. Yes, it only understands English and Danish for now. And yes, it can only return windows/walls/slabs/spaces/doors for now. But imagine if you could ask for "rooms with more than 100 annual hours of overheating"? Or "rooms that are heated by radiators supplied from mixing plant MP02"? The possibilities are endless and I can't wait to start writing some richer queries. Stay tuned and give it a try with your own #BIM model at https://ld-bim.web.app!

Powered by #ifcjs, #nlpjs and #comunica