Recently, me and my 3 future roommates searched the vast web for the prettiest affordable houses to rent in town for next year at university. After over 5 hours of collectively searching we only found 6 places we really like, only to find out a few days later that we will have to look all over once again because another friend will be joining us. All of our university peers spend great amounts of hours searching the web for their next home. I want to save us this time.

What it does

My Project allows users to get the top ten prettiest most affordable house lettings in areas and price-tiers of their chosing within minutes. Where us normal people take hours to skim through hundreds of house listings on a quest to find the perfect home, my application analyses these homes within minutes on the web and gives you a list of the most beautiful and for you affordable homes around.


Using a dataset of various room images with a rating of their beauty and with the aid of Microsoft Azure Custom Vision, I was able to train model to learn to predict how aesthetically pleasing given rooms look. Given this, this application aids users and especially students to easily and quickly find respectable places to rent by web scraping property-sites and applying the machine learning model to evaluate which properties rate the highest, finding the user the most aesthetically pleasing places while still being affordable, saving them a great number of hours searching through the vast web.

As an alternative to housing price prediction models, which do not factor in the aesthetics of properties, this application can also be used by letting landlords to easily explore housing situations in the same area as their own property in a unique way to get a better understanding of how good their property rates against the competition and to uncover whether their selling factors require changes.

Built With

  • microsoft-azure-custom-vision
  • python
  • java
  • datasets
  • webscraping
Share this project: