Welcome to Yillow!
Hi! Welcome to my journey into the world of machine learning and Data Science as I unravel the complexities of the housing market in "Yillow: A Parody of Zillow." In this comprehensive project, there's a lot of things to explore: data science, machine learning model construction, market trends analysis, interactive maps, and a real-world application of my model, the "Yestimate," in finding under-valued gems on the housing market. There's lots to explore, so scroll down to see more information about which section you should start at!
Motivation - And Why Does This Look Like Zillow?
This project stemmed from my deep-seated interest in machine learning/AI, data science, and interest in (fingers crossed) buying a house one day. As a rising senior in college with sights set on becoming an ML/AI/Software Engineer, I was compelled to craft a project that would not only showcase my skills but also provide a meaningful platform for learning and exploration. The thought of building a predictive model for housing prices, inspired by Zillow, struck me as the perfect blend of a formidable technical challenge and a practical, real-world application. So, I decided to create a project in parody-style of the real Zillow, which is why the sites look so similar!
Where to Start?
Here's a brief overview of each section, feel free to check out whichever one seems most interesting to you!
Data Science
Starting with gathering a dataset of ~5,000 houses from Zillow, I embarked on the critical stages of data cleaning and preprocessing. Afterwards, I developed feature engineering methods to increase the information my model would have to work with, and I also explored methods of using convolutional neural networks to analyze features from images of houses.
Data Science
Machine Learning
Machine learning lies at the heart of this project, driving the creation of the Yestimate. Here, we delve into the depths of various models, testing their ability to predict housing prices accurately. Several models are spot-checked, with the most promising ones selected for further fine-tuning and optimization. An ensemble approach is also adopted to enhance performance.
Along with constructing the Yestimate, this section also owns up to the many nuances as to the flaws associated with evaluating the Yestimate against the Zestimate and just how good the Yestimate really is.
Machine Learning
Price Analysis
In this section, I pull back the curtain to reveal the intricate patterns and trends hidden within my dataset. I delve into exactly how bedrooms, bathrooms, and living area affect a house's price, and I also explore which features are most important in determining a house's price. Recognizing that "location, location, location" often rings true in real estate, I examine the role of how geography incluences house price.
Pricing Analysis
Yillow Maps
A central piece of this project is the Yestimate, and my custom-built interactive maps allow you to explore the Yestimate in a fun, engaging way. This is intended to parody Zillow's own maps feature, and you can zoom in, scroll around, and even click on the house dots to get more information about them! This is a must-do before you leave Yillow.
Yillow Maps
Home Hunt
An exciting twist in this project is the Home Hunt. I set out on a virtual quest to find my dream home using the Machine Learning model I developed, my personal preferences, and the expansive dataset at my disposal. This not only serves as a practical demonstration of the model's capabilities but also adds a fun, personal touch to the project. I invite you to follow me on this exciting journey as I hunt for the perfect home and save hundreds of thousands of dollars!
Home Hunt
Owning It
In order to be a true parody-style project of Zillow, I also need to live up to Zillow's values, particularly owning up to what you've done. This comes in two primary sections. First, I have a section where I explain the nuances behind saying how the Yestimate has an error only about twice that of the real Zestimate. This is below, so make sure to check this section out before you go:
Owning ItThe second section where I "Own It" is the summary of what I've done here, which if you're pressed for time, is a good way to quickly see what I accomplished with this project, what I learned, and what I would do differently in the future:
SummaryAs a note, these two pages are probably the most text-heavy pages on all of Yillow, but I think that's necessary in order to Own Up to what I've done in this project.
I hope you enjoy your stay at Yillow!
