Automated Machine Learning (AutoML) Web Application
Utilizing Python, Streamlit, PyCaret, and Pandas Profiling, this project develops an interactive AutoML web application, allowing data upload, exploration, modeling, and download of the best model
This project focuses on the development of a web application for Automated Machine Learning (AutoML). The application allows users to upload datasets, perform exploratory data analysis, build regression models, and download the best model for further use. The application utilizes Streamlit, Plotly, PyCaret, and Pandas Profiling to provide an interactive and streamlined experience for data analysis and modeling.
The Automated Machine Learning (AutoML) Web Application is designed to automate various stages of the machine learning process, including data upload, exploratory data analysis, model building, and model download. The application leverages the capabilities of Streamlit, Plotly, PyCaret, and Pandas Profiling to simplify and expedite the data analysis and modeling workflow.
The web application starts with a sidebar that displays the AutoML logo and provides navigation options for data upload, exploratory data analysis, model building, and model download. Users can choose the desired option based on their requirements.
If the user selects the "Upload" option, they can upload their dataset using the file uploader. The application reads the uploaded dataset into a Pandas DataFrame and saves it as a CSV file for future use. The uploaded dataset is then displayed to the user for verification.
Choosing the "Profiling" option enables users to perform exploratory data analysis on the uploaded dataset. The application utilizes Pandas Profiling to generate a comprehensive profile report that includes statistical insights, data quality checks, variable correlations, and visualizations. The profile report is displayed to the user for better understanding and analysis of the dataset.
Selecting the "Modelling" option allows users to build regression models using the PyCaret library. The user can choose the target column for prediction from a dropdown menu. The application converts string and categorical columns into numeric representation for modeling purposes. Upon clicking the "Run Modeling" button, the application sets up the dataset, pulls the preprocessed data, and compares multiple regression models. The best-performing model is saved and displayed to the user. If the user chooses the "Download" option, they can download the best-performing model as a pickle file for further use and deployment.
Programming Language: Python
Libraries: StreamLit, PyCaret, Pandas Profiling
We can develop projects with similar requirements tailored to your needs, or create custom solutions specific to your requirements. This demo showcases the coding and functionality of the project, and we can customize the user interface (UI) according to your specific requirements. We can also seamlessly integrate this functionality into your existing web or mobile application, ensuring a smooth user experience across platforms.