Choosing your new home is a live changing decision and the emotional component plays a huge part in it. With this is mind, my customer was willing to develop a real state listing match maker the would go far beyond square feets, bedrooms and car parking. Idea was to give the listing emotional characteristics the buyers would identify themselves with.
By leveraging first party owned data as well as third party data such as businesses, events and social media posts as non exhaustive sources of information, I've developed an architecture that leveraged a hand full of Large language Models (LLM) to interpret emotional and lifestyle spontaneous expressions on the neighbourhood, classifying them accordingly to themes like types of events, pets, parks, sports, music and associated emotions to it, so this data could be aggregated to real state listing, making possible to develop a math making algorithm allowing buyers to find the home for life.
