In this individual term project, you'll apply the complete predictive analytics workflow to a real-world dataset. This project simulates professional analytics work and produces a portfolio-ready GitHub repository that you can showcase to potential employers.
This is a 4-week project (November 18 - December 12) designed to be completed with steady, focused work. The timeline below provides helpful checkpoints to guide your progress. With consistent effort of 10-15 hours per week, you'll create something you can be proud of!
📦 What you'll create: A public GitHub repository showcasing your complete term project. Submit the link via Canvas under the 'Term Project' assignment by Friday, December 12th at 11:59 PM.
This repository becomes part of your professional portfolio. You can share it with potential employers during job interviews and internship applications.
Think of your repository as a living showcase of your skills. A well-crafted README with clear visualizations, organized code, and thoughtful documentation tells your story better than slides ever could.
Selected top projects will be featured in:
Professional introduction including:
Analysis notebooks with:
💡 Pro Tip: You can set up your GitHub repository at any time during the project. Many students find it helpful to create it early (Week 1 or 2) and gradually add content as they progress. This approach makes the final week less hectic!
The milestones below are suggested checkpoints to help you pace yourself effectively. They're based on how successful students have structured their work in the past. Feel free to adapt the pace to your schedule, but staying roughly on track will help ensure you have time for quality work and revisions.
Optional: Share a brief update in the Canvas discussion board if you'd like early feedback
Note about Thanksgiving (Nov 28-29): This checkpoint period includes the Thanksgiving holiday. Many students find it helpful to make good progress in Week 1 so they can take a well-deserved break or work at a lighter pace during the holiday week.
Submit via Canvas: GitHub link + PDF report
| Week | Focus Area | Suggested Activities |
|---|---|---|
| Week 1 Nov 18-22 |
Dataset Selection & Initial EDA | Review guidelines, explore potential datasets, select your dataset, complete initial exploratory analysis, draft 2-3 business questions |
| Week 2 Nov 25-29 🦃 Thanksgiving Week |
Data Preprocessing & Cleaning | Clean data, handle missing values, encode features, engineer new features, create train/test split. Many students work ahead before Thanksgiving and take a lighter week during the holiday. |
| Week 3 Dec 2-6 |
Model Building & Evaluation | Implement 2-3 models, compare performance, tune hyperparameters, interpret results, document findings. Start drafting report sections. |
| Week 4 Dec 9-12 |
Final Report & Polishing | Complete final report, organize GitHub repository, polish your README, proofread everything, test all links. Submit by Friday Dec 12th at 11:59 PM. |
What to do:
Dataset Requirements:
Deliverable: Dataset source, link, and brief rationale (1-2 paragraphs) in your report
What to do:
Deliverable in Report: Summary of EDA findings with at least 6 meaningful plots
Deliverable in Colab: Well-documented EDA notebook with clear markdown explanations
What to do:
Deliverable: Clear documentation of preprocessing steps in both code and report
What to do:
Deliverable: Complete model comparison with evaluation metrics, visualizations, and interpretation
💡 About the Model Requirement: The requirement is a minimum of 2 models to ensure you can compare approaches. Many students find it natural to try 3-4 models as they explore different techniques. Quality analysis of 2 well-chosen models is better than superficial implementation of many models!
What to do:
Deliverable: Business-focused interpretation section in report with clear recommendations
What to consider:
Deliverable: 1-2 page ethics reflection in your report
Format: Professional business report (PDF or Markdown), 10-12 pages
Audience: Write for a business-savvy reader who may not know machine learning details but wants to understand your approach and findings
💡 Writing Tips: Start drafting sections as you complete each phase of analysis. For example, write your EDA section right after completing the exploratory analysis while insights are fresh. This makes Week 4 much more manageable!
Total Points: 100
Your project will be evaluated on both technical execution and communication quality. Strong work in both areas creates an excellent portfolio piece!
| Component | Points | Key Evaluation Criteria |
|---|---|---|
| GitHub Repository Quality | 15 | Professional README, organized structure, working Colab links, clear documentation |
| Exploratory Data Analysis | 15 | Comprehensive exploration, meaningful visualizations, insightful observations |
| Data Preprocessing | 15 | Appropriate cleaning, thoughtful feature engineering, well-documented decisions |
| Model Development | 20 | Multiple appropriate models, proper implementation, hyperparameter tuning |
| Model Evaluation | 15 | Appropriate metrics, thorough comparison, sound model selection reasoning |
| Business Insights | 10 | Clear translation to business value, actionable recommendations, practical considerations |
| Ethics & Interpretability | 5 | Thoughtful reflection on bias, fairness, responsible deployment |
| Report Quality | 5 | Professional formatting, clear writing, proper structure, good visualizations |
This is an individual project. You may:
You may NOT:
You're welcome to use AI assistants to help learn concepts and debug code, with these expectations:
💡 AI Tools for Learning: AI assistants work best when you use them to understand concepts, explain errors, or explore alternative approaches - not as a shortcut to skip learning. The goal is to build skills you'll use in your career!
Properly cite all external resources:
These datasets are pre-vetted and ready to use for your project. You can start immediately with any of these - no approval needed! You're also welcome to find your own dataset (just reach out if you'd like early feedback on your choice).
Pick something that interests you! You'll spend several weeks with this data, so choose a topic you find engaging. Healthcare, business, finance, HR - whatever sparks your curiosity will make the project more enjoyable.
Quick tips:
Prefer to explore on your own? These are great places to find datasets:
I'm here to support your success! Don't hesitate to reach out if you have questions or want feedback.
Office Hours: Wednesdays, 2:00 PM - 3:00 PM
🔗 Join Zoom Office Hours
Email: harslan@suffolk.edu
Best Practices for Questions:
LastName_FirstName_ISOM835_Project.pdf💡 Submission Tip: Test your GitHub links in an incognito window to make sure they're publicly accessible. Have a friend or classmate verify they can view your repository!
Late Policy: Please reach out if you anticipate any challenges meeting the deadline. I'm happy to discuss options and support your success.
A: Yes, if you're doing a substantially different analysis with different models and questions. Please check with me first to ensure there's enough new content.
A: Absolutely! This timeline is designed based on successful past projects. With 10-15 hours per week of focused work, you'll have plenty of time for quality analysis. The weekly checkpoints help you pace yourself effectively.
A: Many students find it helpful to make good progress in Week 1 so they can enjoy Thanksgiving break or work at a lighter pace. Others prefer to work steadily through all four weeks. Choose the approach that fits your schedule and energy best!
A: Yes, especially early in the project. If you realize your dataset isn't working well, it's better to switch early (ideally within the first week or two) than to struggle with a problematic dataset. Just let me know if you need to make a change.
A: That's completely okay! Real-world data doesn't always produce perfect predictions. What matters is demonstrating proper methodology, showing what you tried, and thoughtfully explaining the challenges. Honest analysis of limitations shows professional maturity.
A: Both are important! The report demonstrates your understanding and insights; GitHub demonstrates your technical skills and creates a portfolio piece. They complement each other.
A: No problem! Just make sure it's public before the deadline. Double-check the visibility settings when you submit.
A: Write for a business-savvy reader who may not know machine learning details. Include enough technical information to demonstrate your work, but focus on clear explanations and business implications.
A: Yes, comparing different approaches is essential for understanding what works best for your problem. Most students naturally explore 2-3 models as they learn which techniques suit their data. Quality comparison of 2 models is better than superficial implementation of many!
A: Absolutely! Office hours are perfect for getting feedback on your dataset selection, approach, or preliminary results. The earlier you check in, the more time you have to incorporate suggestions.
This project is your opportunity to demonstrate everything you've learned and build a portfolio piece you can be proud of.
Remember: I'm here to support your success. Don't hesitate to reach out during office hours or via email if you have questions or want feedback at any stage.
Looking forward to seeing your excellent work! 🌟