Community Watch

Inspiration

There are a lot of issues when investing in real estate, mainly associated with the research process. Including having to search multiple websites to find relevant information, information being hidden behind paywalls and lack of enjoyable user experience. This all sucks the job out of investing.

What it does

Community Watch fixes those issues by providing an interactive user friendly web app where all the most relevant data real estate agents need to know is displayed. With the ability to save and explore different communities using the interactive map tool, it makes finding the right community to buy real estate more enjoyable.

How we built it

The app was built using ReactJS and implemented the Google Maps APIs. A geojson file of Toronto’s communities was used to draw the various Polygons and compare with our processed data. ChartJS was also used to display the historical price charts and sales data. The data pipeline is built with python pandas to aggregate and sort data. The ML component that outputs safety, transportation and education index is done using a CNN with TF.keras. Most of the data we sourced was from Federal open data sources including: Break and Enters Theft Robbery

Challenges we ran into

There was virtually never the exact data we needed, so we used Machine learning to predict those values using the open sourced data we had combined with some web scraping for target values. Also, we had our first experience with Google Maps API and there were very few docs on the API implementation in React. We needed to check numerous different 3rd party tutorials and try various ways to implement the API. Once we overcame that, we were able to smoothly integrate our other components.

Accomplishments that we're proud of

We are proud that we came together and built a full project despite the barrier of not being able to work in-person. It was difficult to start moving at first, but ultimately we were able to communicate better and work cohesively to build this.

What we learned

We learned that there is a lot of open-sourced data online that we can use in future projects. A large part of this was learning how to search for data and how to find accurate and relevant datasets that pertain to our idea.

What's next for Community Watch

Automate the entire data pipeline and move the backend onto a program such as FireBase. This would clean up the code on the front end and further smoothen out the user experience. We also would like to continue finding data in other regions such as Richmond HIll and Markham and implement all the data together.

Built With

Share this project:

Updates