Princeton Consultants reposted this
Join us for an Analytics+ #2025Analytics Workshop: Data-First Best Practices & Integrations with Princeton Consultants & Databricks with Irv Lustig and Siddhesh Pore Many O.R. and Machine Learning textbooks and guides focus on the models and algorithms. Automated machine learning software assumes that your data is in an easily consumable format. Optimization software providers often give examples assuming that data is organized in data structures appropriate for modeling languages or programming idioms. In the real world, data is messy, and this workshop will address how to handle this challenge and achieve outstanding results. Significant work is required to organize and understand data and to successfully deliver advanced analytics solutions. Irv Lustig will discuss a “Data-First” approach and, using examples, illustrate best practices for organizing data for machine learning and optimization with considerations for deployment in applications. Irv will also discuss the pitfalls that can occur by not taking numerous data-related issues into account when developing an application. Siddhesh (Sid) Pore of Databricks will then discuss how Databricks’s Data Intelligence platform enables building optimization techniques on the platform. Sid will present an overview of Databricks and its ability to handle large-scale data processing and analytics. Sid will delve into the integration of solvers within the Databricks environment, discussing various implementation options and best practices. The core of the discussion will focus on how Databricks can be used to prepare data, build optimization models, and efficiently solve complex decision problems at scale. Sid will cover the process of integrating solvers, managing data flows, and visualizing results. Commercial-grade solvers like Gurobi Optimization, along with data and analytics platforms like Databricks, are increasingly being used by businesses to address optimization challenges. These platforms help prepare data inputs and turn solver outputs into actionable applications. To illustrate these concepts Sid will present a live demonstration, using Gurobi as the solver, walking step by step through integrating solves within Databricks.