HouseCheck

Initial Page
Filling Form
The Magic

Inspiration

While moving to a new city, people spent a lot more effort in finding a good place to rent. Even if they find the good place they remain suspicious whehther all amenities mentioned will be there or not. Although some services provide verified properties but that mainly involves physically visiting the property. So, we designed a verification service which leverages state of the art ML models and detect the amenities available there. For landlords, we made it easy to advertise their property on the platform, they just upload the some pictures and select the amenities they are offering and ChatGPT will give them a catchy advertisement to attract many tenants.

What it does

Our service is mainly a verification service. For tenants the service provide a holistic view of amenities and property type available using computer vision technology. For landlords, it automatically creates an advertisement with catchy title and description by leveraging large language models.

How we built it

Multiple technologies went into creating this amazing view of the service. Backend was designed using Flask framework in python. The frontend was designed in the react which provided a component view of modules. For DevOps docker was used to containerize and ship the whole application at scale capable of catering problems faced in the real world. Powered by restb.ai for providing computer vision solutions as restful service along with large real state data. The real-state tags were fed to ChatGPT for creating fancy advertisement to naturally attract the tenants.

Challenges we ran into

Containerize the whole application was very challenging along with communication between multiple react components. There are some limitation of restb.ai platform for example detecting whether multiple images belong to the same room.

Accomplishments that we're proud of

We are glad that we developed a solution for a problem faced by a lot of people. This was also inspired by our own experience we faced while moving to Barcelona, that's why we tried to develop a more smoother experience for the tenants and the landlords both.

What we learned

It was our first time developing a solution based on computer vision solution so that certainly was a big learning outcome from this hackathon along with integrating the large language model in our application. Being the student of masters in big data management and analytics, we certainly got the hands on experience of building a scalable application.

What's next for HouseCheck

Identifying the images belonging to same room or property even in different angles. Incorporating video analytics to give more immersive and real-time annotation in the video. Creating the whole descriptive advertisement portfolio of property from media only. Creating 3D projection of 2d view points to give the most immersive view of the property for the tenants and remove the possibility of false information.