README checks

[RaymondBrien]

• ML business case section has these sections and uses keywords defined underneath this list: We want ML model to predict... Our ideal outcome is to... The model successs metrics are... The model output is... Heuristics... The training data...

Keywords: labels, target, features, variables, attributes, the learning method, and the machine learning tasks you use, train or fit a model, model output, model metrics and prediction

-_____-

README To-Do List

---

1. Add a Table of Contents

• Purpose: Provides readers quick navigation within the README.
• Action: Add links for each major section, such as Business Context, Dataset, Model Design, etc.

---

2. Expand the Business Context

• Purpose: Emphasize the problem and why it matters for the client, Farmy & Foods.
• Action: Include more details on how powdery mildew affects cherry yields and quality and why automation is crucial.

---

3. Provide Dataset Details

• Purpose: Give more context on the dataset, its origin, and structure.
• Action: Add source link, dataset composition (class balance, image dimensions), and dataset purpose (binary classification for detecting powdery mildew).

---

4. Outline Business Requirements

• Purpose: Clarify the client's specific goals and expectations.
• Action: Explicitly list requirements, such as the need for quick detection, reducing inspection time, and ensuring accurate classifications.

---

5. Add Hypothesis and Validation Techniques

• Purpose: Outline the hypothesis guiding model design and methods for validating it.
• Action: Add sections on visual differences (color, texture) between healthy and infected leaves and discuss planned visual analyses (average image, image montage).

---

6. Detail Model Design and Metrics

• Purpose: Clarify the CNN model’s structure and performance metrics to show project results.
• Action:
  • Include details on model architecture (e.g., layers, activation functions).
  • Mention parameter count (6 million) and training methods.
  • Provide metrics achieved (accuracy, loss, etc.) and add relevant plots.

---

7. Describe the Dashboard Design

• Purpose: Show how the app organizes information and user interactions.
• Action:
  • Describe each dashboard page, its content, and functionality.
  • Specify actions users can take, such as uploading images or visualizing performance metrics.

---

8. Expand Technologies Used Section

• Purpose: Ensure readers understand the stack and tools used for each project stage.
• Action:
  • Break down technologies by use (modeling, deployment, visualization, etc.).
  • Include Heroku and Streamlit details, along with image processing libraries.

---

9. Add Deployment Instructions

• Purpose: Guide readers on deploying the project if desired.
• Action:
  • Outline steps for GitHub and Heroku setup, branch selection, deployment process, and any technical requirements (e.g., Heroku stack compatibility).

---

10. Acknowledge References and Credits

• Purpose: Provide appropriate credit for datasets, inspiration, and guidance.
• Action:
  • Add references to similar projects, resources used (e.g., Streamlit, TensorFlow documentation), and any other acknowledgments.

[RaymondBrien]

Important questions to ask: Do we need ML for the solution Is data available for training the model? If not, how can we collect the data? Does the customer need a dashboard or API endpoint? What does success look like? Break down int Epics/User stories Ethics / Privacy concerns What level of prediction performance is needed? What are the project inputts and intended outputs Does the data suggest a particular model? How will the customer benefit?