Project Workflow:

Step 1: Real-time Data Ingestion with Apache Kafka Streams

1. Launch and stream real-time data from the 'customer_churn.csv' file using Apache Kafka Streams.

Step 2: Data Preprocessing with Machine Learning Libraries

2. Perform necessary data preprocessing using libraries such as Sklearn, PySpark MLib, or PyTorch.

Step 3: Supervised Machine Learning Training

3. Train supervised machine learning models (at least 3 models) on the 'customer_churn.csv' training dataset.

Step 4: Model Serialization and Storage

4. Save the best-performing model in .pkl format.

Step 5: Real-time Prediction using the Trained Model

5. Utilize the prepared, trained, and saved model to predict in real-time whether a customer will leave the institution or not based on the 'new_customers.csv' test data.

Step 6: Results Presentation with Web Application Dashboard

6. Present the results in the form of a web application dashboard.

Step 7: Project Upload to GitHub

7. Upload the entire project to GitHub for collaboration and version control.

Tools and Technologies:

• Libraries: Apache Kafka Streams, PySpark MLib, Sklearn, PyTorch, Pandas, Matplotlib
• Frameworks: Flask, Django
• Languages: Python, Java, JavaScript
• Editors: IntelliJ IDEA, Eclipse, VsCode
• Operating Systems: Unix, MacOS, or Windows

Data Description:

• Name: Name of the latest contact at Company
• Age: Customer Age
• Total_Purchase: Total Ads Purchased
• Account_Manager: Binary 0=No manager, 1= Account manager assigned
• Years: Total Years as a customer
• Num_sites: Number of websites that use the service.
• Onboard_date: Date that the name of the latest contact was onboarded
• Location: Client HQ Address
• Company: Name of Client Company
• Churn: Target (label)

Data Source:

Customer Churn Spark Notebook