House Price Predictor

Regression Algorithm
House Model Weights
Cost Function of Algorithm

Inspiration

With COVID coming to an end, more people are looking to buy and sell houses. Thus, we wanted to create a medium to ease this process to make it easier for both buyers and sellers. For sellers, the code will allow them to price their homes accurately. Sellers can create a budget and anticipate price points of homes based on their wants and needs. Therefore, the home-buying process will be more efficient, less time consuming, and overall tension-free.

What it does

Our Loudoun Housing Price Predictor will take inputs such as bedrooms, bathrooms, square feet, etc., and will use those inputs (along with a trained algorithm from previous data) to calculate the predicted price of the house.

How we built it

We trained our model with information from houses all around Loudoun County in order for our program to run accurately. By using characteristics of homes such as number of bedrooms, bathrooms, square footage, the age of the house, and more, our model weights the effects of the values on known prices, calculates and improves the accuracy of the data. Once the cost of the accuracy is lowest, our weights have been efficiently trained. After we trained our program, we made sure to check if our program was accurate with calculating the prices of the houses. Then, by inputting the values of an unknown price, we used the weights to predict the price based on the inputted values.

Challenges we ran into

Our main challenge that we ran into was trying to interpret what the weights meant. We were able to get our program working, but it took quite some time to figure out what the weights meant. Fortunately, we were able to figure it out and it assisted us in further improving our program.

Accomplishments that we're proud of

A major accomplishment of this code was completing it under a time constraint. While our idea was quite simple, it took us quite a while to figure out how we were going to implement it in a short amount of time. Another accomplishment that we are proud of is using real-life data to train our algorithm. We found information about house prices in Loudoun County, which shows that this data is highly accurate in our area. With this, we can predict house prices correctly in the Loudoun County area. This was something that we found interesting and helpful towards our project, as it is an application to our locality.

What we learned

We learned that giving back to the community is not always about service work, donations, or helping only a specific demographic. There are ways to help the general public make their lives easier, and that’s what we hope to accomplish with our project. Our project, for example, targets home buyers and realtors (anything regarding real estate). This gives back to the community by facilitating the several factors into a house’s price, and gives a ballpark estimate so a potential buyer can have an initial price they would pay. We learned that philanthropy like this is just as impactful to the community as service work and volunteerism.

What's next for Loudoun Housing Price Predictor

To improve upon this code, we want to include a factor that accounts for specific location, as that largely affects the prices of homes. Additionally, we hope to take this code to a larger scale by creating a website specifically targeted to realtors and home owners. This way, large real estate developers (such as Zillow, Redfin, etc) will be able to use our algorithm to set a realistic price tag on a home. We can integrate this with other home information to ultimately make a real estate website, selling homes and creating new opportunities to sell and buy. We would also like to expand our data to train the information to other parts of the world. Because our data only contains Loudoun County home data, we could expand it to include data from all over the nation.