City Recommender
CitySelector-ML uses machine learning to recommend the best city for development based on criteria like population and elevation. The model analyzes city attributes to provide data-driven insights for optimal site selection.
Project Statement
A company aims to build a new development and needs to identify the most suitable city that meets specific criteria such as high population, low elevation, and other relevant parameters. The objective of this project is to develop and utilize a machine-learning predictive model that can analyze various city attributes and recommend the best city globally for the company’s project based on the provided criteria.
Responsibilities
- Pre-processed the .csv dataframe by dropping unnecessary features and encoding categorical data.
- Handled missing values and outliers for optimal regression performance.
- Tuned hyperparameters using RandomizedSearchCV to balance overfitting and underfitting.
- Split data into training and validation sets using cross-validation.
- Trained and predicted using RandomForestRegressor, ensuring suitability based on context.
Responsibilities cont.
- Evaluated model performance using metrics like mean_absolute_error() and r2_score().
- Achieved strong results, confirming the effectiveness of the chosen model and hyperparameters.
- Identified and removed outliers using Z-scores and box plots to improve prediction accuracy.
- Visualized predicted vs. actual points on a 3D globe to assess model accuracy interactively.
Technologies
- Python
- pandas
- numpy
- scipy
- sklearn