maxxies/accra-house-pricing-ml-model

A machine learning model that predicts house prices in Accra # accra-house-pricing-ml-model A machine learning model to predict the price of a house in Accra, Ghana when given the location, number of bedroomws, number of bathrooms and number of garages. ## Table of Contents * Overview * Sections * Overview * Data * Model * Deployment ## Overview This a machine learning model centered on predicting house prices in Accra. This is deployed using a web framework for user interaction with the model. The model makes use of data from a trusted source. ## Sections ### Overview - Building the the predictor makes use of many processes or actions which would be talked about. These have been worked out in different files as seen above. ### Data - The data used for this project was scrapped from a trusted website, meqasa.com. - Selenium and BeautifulSoup were options to be used for the scraping but due to the nature of the website, selenium was used. - Selenium offered web interactions which helped in scrapping the data unlike BeautifulSoup. - Illustrations on using both libraries for scraping the website has been done. - The data comprises of : - Location of the house - Number of available bedrooms - Number of available bathrooms - Number of available garages - Additional information like the longitude and latitude of the location of the house was added to the data with the help of a library in python. - Data was cleaned and fed to the model. ### Model - Consideration were made on predictors like Linear Regression, Decision Tree Regressor, Random Forest Regressor and Support Vector Regressor(SVR). - After data was split into the train, test and validation sets, they were tested against all these predictors to get the one that works best on them. - The model was then built on the chosen predictor after finding the best parameters for it. - The model was then saved to be used by the flask app. ### Deployment - The interface that was used to serve the model to users is a web app built with Flask. - The saved model is used up in the app t …