Getting Started with Mathematical Optimization in Python

By Jerry Yurchisin, Data Science Strategist

As a data scientist, you’re likely familiar with Python—the go-to language for many data science tasks, from data analysis to machine learning. But have you considered using Python for mathematical optimization? With the right tools and resources, you can harness the power of mathematical optimization in Python to solve complex problems and make optimal decisions.

Why Python for Mathematical Optimization?

Python is a versatile, easy-to-learn language with a rich ecosystem of libraries and tools for data scientists. It’s also a great language for mathematical optimization, thanks to libraries like Gurobi that provide powerful optimization solvers with Python interfaces.

Using Python for mathematical optimization offers several benefits:

1. User-friendly syntax: Python’s simple syntax and readability make it easy to formulate and solve optimization models.
2. Fits well with data science workflows: If you’re already using Python for data analysis or machine learning, adding mathematical optimization to your workflow is a natural extension.
3. Access to powerful optimization solvers: Libraries like Gurobi provide access to powerful optimization solvers that can handle large, complex problems.

Learn About Using Gurobi in Python

For a brief introduction to mathematical optimization for data scientists, you may want to start with these on-demand webinars. In the recordings, we introduce our latest Python notebook examples and demonstrate how you can combine machine learning predictions and optimized decision recommendations.

• Where Data Meets Decisions: Part 1Learn the key components in the formulation of mixed-integer programming (MIP) problems. You will learn how to use the Gurobi Optimizer to compute an optimal solution of the MIP model.

  Watch Video

• Where Data Meets Decisions: Part 2The video introduces new educational resources for both instructors and students to use for teaching and learning real-world applications of problems that combine data science and mathematical optimization.

  Watch Video

• Opti101: Optimization Training for Data ScientistsIn this on-demand, self-paced training, optimization experts will guide you through the basics and give you the tools to formulate and code a comprehensive application that will optimize rebalancing of bike-share stations using a machine learning-based forecast.

  Watch the training series, “Opti101.” Our popular Opti201 series will be available on-demand later in 2024.

Jump Into Our Gurobi Jupyter Notebook Examples

The fastest and easiest way to try out Gurobi’s Python interface is through our Jupyter Notebook library. Although we have dozens to choose from, we recommend starting with the following:

• Intro to Mathematical Optimization ModelingLearn the key components in the formulation of mixed-integer programming (MIP) problems. You will learn how to use the Gurobi Optimizer to compute an optimal solution of the MIP model.

  View Model

• Avocado Price OptimizationThey say when life gives you avocados, make guacamole. Just like the perfect guacamole needs the right blend of onion, lemon, and spices, finding an optimal avocado price needs the right blend of descriptive, predictive, and prescriptive analytics.

  View Model

• Best Feature Selection for ForecastingA linear regression problem that minimizes the residual sum of squares subject to the constraint that the number of non-zero feature weights should be less than or equal to a given upper limit.

  View Model

• Combining Machine Learning and Optimization Modeling in Fantasy BasketballUtilizes supervised machine learning to predict basketball players’ fantasy scores from historical data and formulates an integer programming model to build the optimal lineup

  View Model

• Music Recommendation SystemA good song never gets old, except when it does. In this notebook, we will create a music recommendation system using a mixture of predictive and prescriptive analytics.

  View Model

• Text Dissimilarity Using Linear ProgrammingWith text-generation AI becoming popular and ubiquitous, can mathematical optimization catch cases of plagiarism?

  View Model

• Airline Planning After Flight DisruptionWeather events are a major threat to the airline industry. The unpredictable nature of snowstorms, heavy rains, and icy runways make it difficult for aviation planners to make accurate schedules.

  View Model

Expand Your Data Science Skills with Mathematical Optimization

Adding mathematical optimization to your data science skill set can open up new possibilities for problem-solving and decision-making. Whether you’re optimizing supply chains, scheduling resources, or making strategic decisions, mathematical optimization can provide powerful, actionable insights.

And be sure to check out our additional free learning resources to keep your momentum going, including our new Gurobi AI Modeling resource—which features a custom GPT specialized in converting plain-text problem descriptions into mathematical models and Python code using Gurobi.