nach oben

2023 | Buch

Kapitel lesen Erstes Kapitel lesen

Predictive and Simulation Analytics

Deeper Insights for Better Business Decisions

verfasst von: Walter R. Paczkowski

Verlag: Springer International Publishing

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book connects predictive analytics and simulation analytics, with the end goal of providing Rich Information to stakeholders in complex systems to direct data-driven decisions. Readers will explore methods for extracting information from data, work with simple and complex systems, and meld multiple forms of analytics for a more nuanced understanding of data science. The methods can be readily applied to business problems such as demand measurement and forecasting, predictive modeling, pricing analytics including elasticity estimation, customer satisfaction assessment, market research, new product development, and more. The book includes Python examples in Jupyter notebooks, available at the book's affiliated Github.

This volume is intended for current and aspiring business data analysts, data scientists, and market research professionals, in both the private and public sectors.

Inhaltsverzeichnis

Frontmatter

The Analytics Quest: The Drive for Rich Information

Frontmatter

Chapter 1. Decisions, Information, and Data

Abstract

Know your audience is a well-known advice often quoted in public speaking, effective presentation, or creative writing courses. You are then taught to target their interests and concerns which are sometimes called pain points. Your employer or client, perhaps a CEO, is your audience. What are her pain points? What keeps her up at night? How does she address and manage them? Basically, how does she manage her business?

Walter R. Paczkowski

Chapter 2. A Systems Perspective

Abstract

One of the four themes of this book is that information is buried inside data: data per se are not information, but contain information which must be extracted. There are many extraction methods that can be grouped into classes as I have shown in Fig. 2.1. The two umbrella classes are Business Intelligence and Business Analytics. The former tells a decision maker what did or is currently happening. The latter tells him/her what will happen and, possibly, what should be done. These two classes are umbrella classes because each has more specific subclasses.

Walter R. Paczkowski

Predictive Analytics: Background

Frontmatter

Chapter 3. Information Extraction: Basic Time Series Methods

Abstract

All data share a common structure, and all models are meant to reflect that data share the same structure.

Walter R. Paczkowski

Chapter 4. Information Extraction: Advanced Time Series Methods

Abstract

I developed a few commonly used prediction methods in Chap. 3, methods that rely on time series data. For many situations, primarily what I will later call operational scale-views (the day-to-day operations of a business that are narrowly focused), they are sufficient. There are, however, more complex situations for which these methods are insufficient. The data and problems are more intricate and often convoluted, and thus the prediction methods must also be more intricate. This will be the case for what I will later call tactical scale-views. The complexity of the data and problems dictate the methods. I will develop more intricate models in this chapter, models that still rely on time series data, but just more of it.

Walter R. Paczkowski

Chapter 5. Information Extraction: Non-time Series Methods

Abstract

I focused on time series data for predicting an outcome in Chaps. 3 and 4.

Walter R. Paczkowski

Chapter 6. Useful Life of a Predictive Model

Abstract

The material I covered in Chaps. 3 –5 only skims the surface of available predictive methods. There are many more than it is possible to cover in this book. The available methods cover a wide range of prediction topics and problems so that, as a collective whole, their scope is very broad. Because of this, each is worthy of a separate book. Also, the depth of each is deep so my coverage only provided the highlights of each of those I addressed. As with the scope, full coverage of each method requires a separate book.

Walter R. Paczkowski

Simulation Analytics: Background

Frontmatter

Chapter 7. Introduction to Simulations

Abstract

Simulations comprise the latest stage in what I will refer to as the Science-Technology Revolutionary Period. This is the result of a gradual but definite evolution in our studying of how the world works that began in the late sixteenth century with Copernicus in 1543. Certainly, the Ancient Greeks, most notably Aristotle, and the Ancient Romans pondered the big questions: the nature of life, nature itself, and the universe. Their simplistic views held sway for centuries until the Scientific Revolution of the sixteenth century that marks the beginning of the period. It was then that our methods for studying the “big questions” led to discoveries and technologies.

Walter R. Paczkowski

Chapter 8. Designing and Analyzing a Simulation

Abstract

I noted in Chap. 7 that there are two general application domains for simulations: scientific research and practical problems where the latter could be business or public policy.

Walter R. Paczkowski

Chapter 9. Random Numbers: The Backbone of Stochastic Simulations

Abstract

I outlined the general process of simulation design in Chap. 8. One topic I did not address applies to stochastic simulations. This is the generation of random occurrences based on random numbers. Random numbers are what make simulations stochastic. In fact, “stochastic” means random. But what is a random number and how are they generated? These are the questions for this chapter.

Walter R. Paczkowski

Chapter 10. Examples of Stochastic Simulations: Monte Carlo Simulations

Abstract

I introduced random numbers in Chap. 9 as the backbone for stochastic simulations. Simulations are one part of a two-part process for analyzing and predicting consequences, what I called these implications and ramifications (I&R), of business decisions. These are usually broader in scope than typically assumed. The reason for this broad scope is the complex system nature of a business. Because of this complexity, an impact on one part of the system affects all the other parts. The only way to understand the I&R for the entire system is via simulations. One very important example of a simulation is the Monte Carlo simulation, which is my subject for this chapter.

Walter R. Paczkowski

Melding the Two Analytics

Frontmatter

Chapter 11. Melding Predictive and Simulation Analytics

Abstract

I introduce the melding of Predictive and Simulation Analytics in this chapter. This is best done by examples. The perspective for the examples varies by the scale view of a decision-maker. Since the scale views can be many and varied, I chose to use the three I frequently mentioned in earlier chapters: operational, tactical, and strategic. These are only meant to illustrate my ideas.

Walter R. Paczkowski

Chapter 12. Applications: Operational Scale View

Abstract

This is the first of three chapters that contain examples of the melding of predictive and simulation analytics. The perspective for these examples varies by the scale view of a decision-maker. Since the scale views can be many and varied, I chose to use the three I mentioned many times in earlier chapters: operational, tactical, and strategic. This chapter focuses on the operational. The examples are only meant to illustrate my ideas. For each one, I will begin with a discussion of a simulation and then introduce predictions into the simulation framework.

Walter R. Paczkowski

Chapter 13. Applications: Tactical and Strategic Scale Views

Abstract

This chapter continues the examples of the melding of predictive and simulation analytics with a focus on a tactical and strategic scale view. The creation of predictions is the same as for the operational scale view, but the data and the nature of the model change as you might expect. The simulations, however, could potentially be different. Different how? Operational scale view simulations are predominantly stochastic. Both the tactical and strategic scale view simulations could be stochastic or non-stochastic. If the latter, then they are in the class of system dynamics (SD) simulations with what-if scenarios. It all depends on the problem.

Walter R. Paczkowski

Backmatter

Titel: Predictive and Simulation Analytics
verfasst von: Walter R. Paczkowski
Verlag: Springer International Publishing
Electronic ISBN: 978-3-031-31887-0
Print ISBN: 978-3-031-31886-3
DOI: https://doi.org/10.1007/978-3-031-31887-0