This project is about creating an app using machine learning approch , which predicts the price of the flight using the inputs like the departure date and time , arrival date and time , source city and the destination city , the number of stops and finally the airline in which they would like to travel. Using all these features the price of the flight will be predicted.
- Data Exploration : Getting an idea about the data , type of features , number of categorical and numerica variables and creating plots to get a better understanding of the data.
- Feature Engineering : Converting the categorical variables into numeric variables using One hot encoding and Label Encoder.
- Feature Scaling : Transforming the data into Gaussian Normal Distribution.
- Feature Selection : Selecting the important features and discarding the features which have a high VIF(Variance Inflation Factor) value.
- Model Training and Testing : Multiple regression models are being built using different ML algorithms and the one with best accuracy is selected.
- Hyperparameter Tunning : Random Forest Regressor was selected for predicting the outcome , and it was tunned using differnt parameters to get a better accuracy.
- Web App Development : A web app is made using flask ,python and gunicorn.
- Deployment : Finally the app is being deployed on multiple cloud platforms.
Flight Fare Prediction Web App : (https://flightfarepredicton.herokuapp.com/)
- python
- sklearn
- flask
- html
- css
- bootstrap
- pandas
- numpy
- matplotlib
- seaborn
- gunicorn
- heroku