MGMT 47400: Predictive Analytics

Mitch. Daniels School of Business, Purdue University (Summer 2026 - 4-Week Online Intensive)


IMPORTANT

This document does not replace the official material in the course brightspace page


Course Description and Objectives

This 4-week fully online intensive course enables students to navigate the entire predictive analytics pipeline skillfully—from data preparation and exploration to modeling, assessment, and interpretation. The course runs over 20 business days (May 18 - June 14, 2027) with 112.5 minutes of daily engagement combining short micro-videos (≤12 minutes each), hands-on Google Colab notebooks, exercises, quizzes, and project work.

Students engage with real-world examples through interactive Jupyter notebooks, all designed to run in Google Colab with Google Gemini AI assistance. The course emphasizes essential programming and analytical skills through a “vibe coding” approach (draft → verify → document). By exploring topics such as exploratory data analysis, train/validation/test splits, linear and logistic regression, classification metrics, resampling methods, regularization techniques, tree-based approaches, gradient boosting, model interpretation, and deployment thinking, participants gain both theoretical understanding and practical experience.

The course centers on one comprehensive capstone project that progresses through four weekly milestones: proposal (Day 5), baseline model (Day 10), improved model with interpretation (Day 15), and final executive-ready deliverable including slide narrative and conference-style video (Day 20).

Ultimately, students will leave the course equipped to apply predictive models to data-driven problems, communicate their findings to diverse audiences, and critically evaluate model performance to inform strategic decision-making across various business contexts.

Course Website: https://davi-moreira.github.io/2026Summer_predictive_analytics_purdue_MGMT474/

Instructor: Professor Davi Moreira

  • Email: dmoreira@purdue.edu
  • Office: Young Hall 1007
  • Virtual Office hours: Zoom link in your Course Brightspace Page
  • Individual Appointments: Book time with me through the link in the course syllabus on your Course Brightspace Page or by appointment.

Learning Outcomes

By the conclusion of this 4-week intensive course, students will be able to:

  1. Explain Core Predictive Analytics Concepts: Articulate key principles of statistical learning and predictive analytics, including fundamental terminology, modeling strategies, the role of data-driven insights in business contexts, and deployment considerations—all within a compressed timeline requiring efficient learning and application.

  2. Prepare and Explore Data Effectively: Demonstrate proficiency in cleaning, organizing, and exploring datasets, applying tools and techniques for data preprocessing, feature engineering, exploratory analysis, and preventing data leakage in production-ready pipelines.

  3. Implement Diverse Modeling Techniques: Construct predictive models using linear and logistic regression, classification methods, resampling procedures, regularization techniques (Ridge/Lasso), decision trees, random forests, and gradient boosting.

  4. Assess and Interpret Model Performance: Evaluate the accuracy, robustness, and interpretability of predictive models, critically examining issues such as overfitting, bias-variance trade-offs, cross-validation results, threshold selection, calibration, and fairness considerations.

  5. Communicate Analytical Findings: Present analytical outcomes and model interpretations to technical and non-technical audiences, crafting clear, concise, and visually effective executive-ready deliverables including slide narratives and conference-style videos.

  6. Integrate Predictive Analytics into Decision-Making: Recommend actionable strategies based on model findings, demonstrating the ability to align analytical results with organizational objectives, business costs, and inform evidence-based decision processes with appropriate risk documentation.

Course Materials

  • Textbooks (Required): ISLP James, G., Witten, D., Hastie, T., & Tibshirani, R. (2023). An Introduction to Statistical Learning with Applications in Python. Springer. https://doi.org/10.1007/978-1-0716-2926-2. Download here: https://www.statlearning.com/

  • Computing (Required): A laptop or desktop with internet access and the capability to run Python code through Google Colab: https://colab.research.google.com/.

  • Software (Required):

    • Google Colab is a cloud-based platform that requires no software installation on your local machine; it is accessible through a modern web browser such as Google Chrome, Mozilla Firefox, Microsoft Edge, or Safari. To use Google Colab, you need a Google account and a stable internet connection. All course notebooks are designed to run directly in Google Colab with all necessary dependencies pre-configured. While optional, having tools like a local Python installation (e.g., Anaconda) or a Python IDE (e.g., Jupyter Notebook or VS Code) can be helpful for offline development. Additionally, browser extensions, such as those for VS Code integration, can enhance your experience but are not required. This makes Google Colab convenient and easy for Python programming and data science tasks.
    • Google Gemini in Colab: Students will use Google Gemini AI assistance directly within Colab notebooks to accelerate coding while maintaining accountability through the “vibe coding” workflow: draft code with AI assistance → verify correctness → document decisions. This approach helps students learn faster while developing critical thinking about AI-generated code.
    • Microsoft Copilot: is an AI-powered assistant designed to enhance productivity and streamline workflows across various applications and services. It utilizes large language models and is integrated within Microsoft 365 apps like Word, Excel, PowerPoint, Outlook, and Teams, providing real-time, context-aware assistance for tasks such as drafting documents, analyzing data, managing projects, and communicating more efficiently. Users can leverage Copilot to automate repetitive tasks, generate ideas, summarize information, and access data across their work environment and the web, all within a secure and privacy-conscious framework.

Course Infra-structure

Brightspace: The Course Brightspace Page https://purdue.brightspace.com/ should be checked on a regular basis for announcements and course material.