Rapidly increasing housing prices have created an affordability crisis in Boston. Low-income people and students are suffering the most from this issue, especially month-to-month tenants, as there is currently no rent control in Massachusetts.
What it does
The algorithm takes user inputs for the first and last months of 2023 that the user will be renting. It then outputs the predicted cost at the beginning and end of that period, as well as the average rent during the whole period. These metrics are pivotal to understanding how rent prices can change over a leasing period, and where the market might shift unexpectedly.
How we built it
This model helps inform users about the 2023 renting period by predicting rent prices for different zip codes in Boston. This is done by implementing a linear regression model which utilizes monthly rental data from the past seven years.
Challenges we ran into
Throughout the weekend, we made many changes to our project to ensure the most effective and straightforward implementation of our code. After discovering that our Zillow dataset wasn't suitable for an advanced machine learning model, we decided to employ linear regression instead. Making this change was necessary to maintain more accurate predictions for users, whose financial futures may be dependent on their understanding of local housing markets.
Accomplishments that we're proud of
We are proud of how we were able to adapt after running into problems with the initial plan for our project.
What we learned
This weekend we gained more insight into using and manipulating dataframes to create models and data visualizations. Even though we didn't include machine learning in our final model, we learned more about their functions and when it's best to use them.
What's next for Boston Housing Analysis
Since we lost some time in the beginning trying to use a different model, we were unable to predict as far as we wanted in the linear regression. For the future, we would increase the possible leasing period and output a ranking of the least expensive zip codes, or possibly allow the user to find data for specific months.
Log in or sign up for Devpost to join the conversation.