Objective:
In this workshop, you will apply regression and classification algorithms to predict equipment usage patterns and identify maintenance needs in a FabLab.

Scenario:

FabLabs rely on various types of equipment, such as 3D printers, CNC machines, and laser cutters. Each machine logs usage data, including operating hours, frequency of use, and the number of errors or maintenance issues. Your task is to build a machine learning model to help the FabLab team:

• Predict the operating hours of each machine based on historical data (regression task).
• Classify machines into high-maintenance or low-maintenance categories based on usage and error logs (classification task).

Steps:

1. Data Preparation:

   You are provided with a dataset that contains:

   • Machine type (e.g., 3D printer, CNC, laser cutter)
   • Daily operating hours
   • Frequency of errors
   • Time since last maintenance
   • Total operating hours since installation

   Start by cleaning and preprocessing the data. Handle missing values, encode categorical features, and scale numerical data.

2. Regression Task:

   Use regression algorithms like Linear Regression or Random Forest Regressor to predict the daily operating hours of each machine based on its type, usage patterns, and maintenance history. Evaluate the model using metrics like mean squared error (MSE) or R² score.

3. Classification Task:

   Define thresholds for "high-maintenance" and "low-maintenance" machines based on error frequency and operating hours. Train classification algorithms like Logistic Regression, K-Nearest Neighbors, or Support Vector Machines to classify machines into these categories. Evaluate the model using metrics like accuracy, precision, recall, and F1 score.

4. Model Comparison and Optimization:

   Compare the performance of different regression and classification models. Optimize the models using techniques such as hyperparameter tuning or cross-validation.

5. Insights and Recommendations:

   Use the models to identify machines that are at risk of high maintenance and suggest an optimized maintenance schedule. Predict future operating hours for new equipment and provide recommendations for efficient resource allocation in the FabLab.

Deliverables:

• Jupyter Notebook with:
  • Data preprocessing steps
  • Regression and classification model implementation
  • Model evaluation metrics and comparisons
• A brief report summarizing:
  • Insights gained from the predictions
  • Recommendations for improving equipment maintenance and usage

This workshop integrates real-world data analysis and predictive modeling to support the efficient operation of a FabLab, showcasing the practical application of regression and classification algorithms.

Duration: 40-50 minutes

Objectives:

• Introduce the workshop and its relevance to FabLab operations.
• Explain the dataset and its components (machine types, usage hours, errors, etc.).
• Perform initial data preprocessing:
  • Load the dataset.
  • Encode categorical variables.
  • Split data into training and test sets.

Activity: Students will execute the code for loading and preprocessing data, ensuring they understand each step.

Session 2: Regression Task

Duration: 40-50 minutes

Objectives:

• Introduce regression concepts.
• Build and train a regression model to predict machine operating hours.
• Evaluate the model using MSE and R² score.

Activity: Students implement the regression model, interpret the metrics, and discuss accuracy.

Session 3: Classification Task

Duration: 40-50 minutes

Objectives:

• Explain classification concepts and their application (high vs. low-maintenance machines).
• Train a logistic regression classification model.
• Evaluate the model using accuracy and classification metrics.

Activity: Students will implement the classification model and analyze the results.

Session 4: Results Visualization

Duration: 40-50 minutes

Objectives:

• Visualize regression results (scatter plot of predicted vs. actual values).
• Analyze classification results with a confusion matrix.

Activity: Students generate visualizations and discuss their implications in a FabLab environment.

Session 5: Summary and Recommendations

Duration: 40-50 minutes

Objectives:

• Summarize workshop learnings.
• Interpret model results and propose recommendations for FabLab optimization.

Activity: Students work in groups to prepare a short report with insights and maintenance recommendations based on the models.

This structured session plan ensures students engage with both theoretical and practical aspects of machine learning, applying them effectively in a FabLab context.

Link to jupyter notebook with all the code needed for the workshop: Workshop Code