Real estate business needs fast and efficient tools to take advantageous decisions. Automatic price estimator is a handy tool that can boost the work productivity of real estate projects and navigate the clients in the sea of real estate offers.
The immo-eliza-ml program is an example of such a tool. With the help of such features as location and living area of the apartment of the house it can rapidly predict the price of the property.
To use the model, please check the requirements.txt file. Necessary libraries include Scikit-learn, pandas, numpy, pickle.
To make price predictions, use predict.py. Please specify the folder: houses or apartments, depending on what type of properties you would like to evaluate. Please check the format of the input dataset to carry out the prediction. The predictions of the prices will be written to the y_predict.csv in the folder specified earlier.
To train the model, use train.py. Type houses or apartments depending on your dataset. The model's .pickle file will be saved in the models subfolder, along with the txt file containing the train and the test scores of the model.
Example structure of the folder with input and output files
├───houses
│ ├───data
│ │ dataset.csv
│ │ X_test.csv
│ │ X_train.csv
│ │ y_predict.csv
│ │ y_test.csv
│ │ y_train.csv
│ │
│ ├───models
│ │ ridge.pickle
│ │
│ └───preprocessings
│ preprocessings.pickle
The project was carried out within the framework of Data&AI training by BeCode within 6 days.
- Day 1-2 data preprocessing
- Day 3 model training
- Day 4 model evaluation
- Day 5 structuring
- Day 6 project finalization
The program was developed by Alfiya Khabibullina under the supervision of the coach Vanessa Rivera-Quinones