Hi Simran, Thank you very much for that comprehensive reply!
I will check out all your links. A course on modelling is a good idea as ArangodDB is multi-model and therefore naturally presents interesting new use cases and applications. Your thoughts on my modelling ideas clarify why I got a bit stuck on how to model this scenario as it’s not really a natural fit. I totally agree that price makes more sense as a vertex attribute. However in the context of a recommendation engine I saw some models online where the attributes are vertexes with weights on the edges and thought there might be some clever model that would work, but maybe this doesn’t work with prices and locations as these are more factual properties and can’t/shouldn't really be weighted. As you point out searching users shouldn’t be connected to apartments. Maybe your ideas for a fuzzy search is better where I i.e. increase the price range if a user is less interested in price or vice versa increase the radius if price is more important. You are correct that locations aren’t really point based. Apartments would belong to a polygon district or just a simple tag/category. In theory if a location is close to another district I could also add it to another district but with a lower weight, but then I might as well just use point based search. I guess my use case probably makes more sense to use a graph based recommendations engine if I start using views, matches etc or types of apartments/features to base my recommendations on? Thanks once again for your feedback! Michael -- You received this message because you are subscribed to the Google Groups "ArangoDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to arangodb+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
