-
-
Our Flask Website
Inspiration
Buying a house is the first symbol of financial success in many people's lives. However, as society faces economic hardships, it is more difficult than ever for the first part of the American Dream to be fulfilled, especially for the needs of these people. As a result, we have been inspired to develop an app that has the potential to change the course of the lives of people in the Chicagoland area. By calculating the potential price of a house befitting the demands of specific individuals, we can ensure the creation of appropriate financial goals and support the pathway to financial success.
What it does
Our Flask web app utilizes a machine learning model trained on the Chicago House Price dataset to predict the potential price of a house meeting specific criteria. The app, designed by Aakash Kolli and Akshath Sivachidhambaram, incorporates a user-friendly interface with Python Flask, HTML, and CSS on the front end.
How we built it
Our journey began by importing essential libraries, including Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn, and Pickle, forming the foundation of our comprehensive data science workflow. We meticulously read and preprocessed the dataset, ensuring optimal efficiency and representativeness through techniques like subsampling and missing value imputation.
Challenges we ran into
Throughout the process, we faced challenges in refining the raw data for modeling, selecting the most appropriate machine learning models, and optimizing the app's functionality. Overcoming these hurdles required creative problem-solving and collaboration.
Accomplishments that we're proud of
We take pride in successfully training a Linear Regression model and incorporating a Gradient Boosting Regressor to capture complex relationships within the data. Achieving accurate model evaluations based on metrics like R-squared score and Mean Squared Error showcased the effectiveness of our predictive models.
What we learned
Our journey taught us the importance of meticulous data preprocessing, effective model training, and the significance of visualizations in guiding decision-making. We gained valuable insights into feature interactions and model performance evaluation, enhancing our overall data science skill set.
What's next for House Price Prediction
Looking ahead, our vision involves continuous refinement of the app's features and functionalities. We plan to explore additional machine learning techniques, expand the dataset for more robust predictions, and potentially integrate real-time data for dynamic pricing estimates. The goal is to enhance the app's accuracy and usability, making it a valuable tool for users in the real estate domain and for prospective home buyers.
Our journey in the APEERS high-school hackathon culminated in a locally deployed Flask web app that not only predicts house prices effectively but also represents a significant step forward in our ongoing exploration of data science applications.
Log in or sign up for Devpost to join the conversation.