This is a project that is meant to learn best in industry practices about Linear Regression

• We Preprocess by dropping irrelevant columns, and fill the na values where possible and drop some na values
• We clean the columns making the values as same represenation and casting them into required datatypes
• Add some columns and grouping some rows for better prediction
• Removing outliers in data using common and domain knowledge
• Using dummies to do one hot encoding of categorical column 'location'
• Make the model

• The server is made using flask in python (The models are not copied again in the server folder)
• The server has a main page named server.py that redirects the requests and the prediction and other data processing id done in the util.py page
• '/predict_home_price' is the route that recives the requests and returns a json with the key 'estimated price'
• We also have the /get_loation_names' that returns the location names with key value 'location' which can be used to be displayed in the form as drop down
• the util has two methods one loads the artifacts i.e the odel and column names(column names are necessary while predicting as our model has dummies)
• The other method predicts the estimated price

• The client side is also made using flask
• The app.py redirects requests, we send the variable pred_price to the html page usign jinja2 of flask
• The html page contains the form which when submitted calls '/send_data' routine
• The send_data collects the form data converts it to dictionary and sends it to the flask server usign requests.post method
• The app first loads the location name and the location names are added using jinja template

• Making the UI better
• Deploying on AWS

The project is a learning project.The model and server is inspired by the Youtube tutorial. The client side was implemented using javascript in the tutorial, which I implemented in flask instead to keep the entire project in python.