We are Nomad Agency, a team of five Data Analysts, and we present a web application project that we developed to deploy an interactive Machine Learning model through Dash. This project integrates several tools and libraries within a Python environment, with a focus on the user interface and predictive interactivity.
The goal of this application is to provide a simple user interface to interact with a real-time prediction model based on a dataset. This project leverages visualization libraries like Plotly and uses Dash to structure the interface and manage interactive callbacks.
- Dash: to build the interactive web application.
- Plotly: to create dynamic data visualizations.
- Pandas and Numpy: for data manipulation.
- Scikit-learn: for predictive modeling (classification, regression, etc.).
- Pickle: to serialize and deserialize Machine Learning models.
- Cachetools: to manage caches, optimizing the performance of frequent model calls.
- Gunicorn: for deploying the application on Render.com.
- Data Visualization: Users can explore the data used to train the model through interactive charts.
- Dynamic Predictions: A form allows users to input specific data and instantly receive a model prediction.
- Performance Optimization: Through cache management, response times for frequent queries are reduced.
- Deployment on Render.com: The application is publicly accessible, facilitating demonstration and interaction.
Here is a screenshot of our application's interface before entering any data:
Here is a screenshot of our application's interface after submitting data and displaying the results:
You can explore the live project via this link: Nomad Machine Learning Dash App.
- Security: Future implementation of authorization and authentication management to restrict access to certain parts of the application.
- Scalability: Testing other deployment services and configuring the project to support a higher number of concurrent users.
This project complements the analysis conducted as part of another project:
As part of our Data Analyst Bootcamp, we collaborated as Nomad Agency to explore and analyze a dataset provided by a client. Our mission was to address a commercial operations optimization problem.
To understand the key factors behind the success of a commercial operation, in terms of customer acquisition and new customer recruitment.
To achieve this, we worked with multiple files containing crucial information on products, sales, advertising investments, and user interactions. By analyzing this data, we identified the factors influencing operational performance and proposed recommendations based on our findings.