Data Science2021Data Scientist
King County House Sales
Linear to polynomial regression, 75% → 94%
Exploratory analysis and house-price prediction for King County, USA. A multiple-linear model reached 75% accuracy; train / cross-validation / test splits and a move to polynomial regression lifted that to 94%.
75%Linear
94%Polynomial
King CountyRegion
The Problem & Engineering Constraint
The Core Challenge
A linear price model underfit the King County market. Needed a validated path from EDA to a stronger regressor without leaking the test set.
Technical Architecture & Approach
Engineering Solution & Implementation
Pandas / NumPy / Beautiful Soup for acquisition and EDA, scikit-learn for multiple-linear then polynomial regression with proper splits.
Measured Production Impact
Verified Outcomes & Deliverables
Baseline multiple-linear model at 75% accuracy.
Polynomial model at 94% after train / CV / test discipline.
Technologies & Components
System Tooling & Technologies
PythonPandasNumPyScikit-LearnBeautiful Soup