Scientific Computing Simulations and ModelingTextbook

Time Series Analysis for Data Scientists (Forgy)

Not rated yetRate this book

Learn to examine dependence, stationarity and recurring patterns before modeling temporal data. Forgy’s text develops the mathematical background for autocorrelation, decomposition and ARMA models while distinguishing analysis of past behavior from forecasting future values.

FormatExternal access
LanguageEnglish
SubjectScientific Computing Simulations and Modeling
LevelNot specified
No. of pages—
PublishedNot specified

Read on the publisher’s website. Opens in a new tab.

Time Series Analysis for Data Scientists (Forgy)Free

About this book

Charles Forgy’s Time Series Analysis for Data Scientists approaches observations through their ordering in time and the information that ordering can carry. The introduction distinguishes a formal time-indexed sequence from the practical analytical interest in dependence between observations. Although a time series can contain many types of observations, the text concentrates on numeric data or data represented numerically.

The main sequence prepares the mathematical background, develops autocovariance and autocorrelation, and examines stationarity. It then introduces decomposition and ARMA models. This order encourages readers to investigate a series before choosing a forecasting model: trends, seasonal patterns and changing behavior affect the meaning of later calculations. The book is useful for data science learners who already have some statistical preparation and want a more structured approach to temporal data. The chapter guide emphasizes the stated purposes of each module and includes original exercises for connecting those ideas with a small, clearly documented series.

AI-assisted book summary

Reading guide

Read the introduction first, then review any unfamiliar mathematics. Move through autocorrelation and stationarity before decomposition and ARMA models. Keep a single example series throughout so you can compare how each method changes your understanding of it.

Chapter-by-chapter study guide

1: Introduction

The opening defines a time series as observations ordered by time and notes that the observations need not initially be numeric. The text narrows its practical focus to numeric representations. It also introduces dependence across observations as a central analytical consideration.

Study task: Contrast a sequence of independent coin tosses with its running total and discuss what ordering reveals.

2: Background Math

This background chapter identifies geometric series, complex numbers, variance, covariance and information criteria as preparation for later work. Its introduction invites readers already familiar with these topics to return as needed. It serves as a mathematical reference module within the larger progression.

Study task: Make a prerequisite checklist and work one example involving covariance before continuing.

3: Autocovariance and Autocorrelation

The chapter introduces autocovariance as a central tool and also covers autocorrelation and cross-correlation. Its stated scope includes synthetic and actual time series. This shifts attention from isolated values to relationships between observations separated in time.

Study task: Describe what a lag means for daily data and compare the original series with a one-day shift.

4: Methods to Enforce Stationarity

The introduction identifies trends, unit roots and seasonal fluctuations as reasons a series may be non-stationary. It pairs those issues with detrending, differencing and seasonal adjustment. The chapter also introduces the backshift notation used in subsequent modeling.

Study task: Choose a visibly changing series and justify which feature a proposed transformation is intended to remove.

5: Time Series Decomposition

Decomposition is presented primarily as a tool for understanding historical activity, including long-term trends and seasonal effects. The introduction highlights STL as a useful method. It also distinguishes understanding past patterns from using decomposition within a larger forecasting procedure.

Study task: Sketch a series as trend, seasonal pattern and remainder, then describe an uncertainty in that separation.

6: ARMA Models

The introduction presents autoregressive and moving-average models as foundations for traditional time series work. Autoregression is described through dependence on previous observations. The chapter brings the earlier analysis tools into a framework for both understanding processes and forecasting.

Study task: Explain in words the difference between using earlier observations and earlier shocks as model inputs.

7: Appendices

The source includes an appendices section after the main modeling chapters. This guide does not assign it an unverified substantive chapter summary. Use the linked original to inspect any available supporting material when a main-text calculation or notation requires clarification.

Study task: List one unresolved notation question and locate supporting material in the original resource.

How to study this book

Record the observation interval, units and missing values before calculating anything. Compare an original plot with any transformed series. Explain what each transformation is intended to address and evaluate a model on later observations rather than a randomly shuffled split.

Reflection questions

  • When does the ordering of observations add useful information?
  • Which pattern motivates detrending rather than differencing?
  • What can a decomposition explain without making a forecast?

Sources and coverage

Original book and chapter navigation on LibreTexts

This original guide is based on the verified table of contents and selected opening text from the main chapters, not a complete reading of every chapter, proof, example or exercise. Appendix entries describe available reference resources; they do not claim full appendix review. Mathematical expressions and figures should be checked in the original, since PDF text extraction does not preserve every symbol or diagram.

About access & copyright

This page provides book information and study material. Full-text access is provided by the original publisher or authorised source; a listing does not grant permission to republish the book.

Read our Library Access & Copyright policy

Dependence

Temporal order can reveal relationships between observations.

Stationarity

Trends, unit roots and seasonality motivate different treatments.

Modeling

Decomposition and ARMA address different analytical purposes.

Reader ratings & reviews

Be the first to rate

No ratings yet

0%0%
0%0%
0%0%
0%0%
0%0%

Share your thoughts

Your review helps other students choose the right resources.

Rate this book

How helpful was this book?No sign-up needed. Your rating appears instantly.