Skip to content
ISBN: B0HRD3MV35
ISBN13: 9798141932519

From Past Data to Future Estimates: A Hands-On Introduction to sktime, Time Series Forecasting, and Measuring Prediction Error

$24.99List Price: $24.99
ISBN: B0HRD3MV35
ISBN13: 9798141932519
Formats

Conditions

A brand-new, unused copy in perfect condition.

Formats

Conditions

A brand-new, unused copy in perfect condition.

See all 1 editionsfrom $24.99

Book Overview

From Past Data to Future Estimates

Learn to turn historical data into practical forecasts using Python, sktime, and proven methods for measuring prediction accuracy.

How can you use past observations to estimate what might happen next? How do you know whether a forecast is reliable? And how can you compare different prediction methods to determine which one performs better?

From Past Data to Future Estimates provides a clear, hands-on introduction to time series forecasting for beginners who want to understand the process, write working Python code, and evaluate their results without getting lost in complicated mathematics.

Using sktime, pandas, and matplotlib, this practical guide takes you from understanding historical observations to building forecasting models, measuring prediction errors, and creating a reusable forecasting workflow.

Rather than presenting disconnected examples, the book follows a fictional small business that records monthly customer orders. You'll work with this dataset throughout the lessons, gradually building the skills needed to transform historical records into meaningful future estimates.

What You'll Learn
  • Understand time series fundamentals: Recognize chronological patterns, trends, seasonality, data frequency, and forecasting horizons.

  • Prepare your Python environment: Install and configure Python, sktime, pandas, and matplotlib for practical forecasting projects.

  • Explore and clean historical data: Load CSV files, organize time indexes, identify missing observations, and prepare consistent datasets.

  • Build your first forecasting models: Use sktime to create baseline forecasts and generate predictions for upcoming periods.

  • Evaluate prediction accuracy: Calculate and interpret Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE).

  • Compare forecasting methods: Apply chronological training and testing, examine alternative models, and use prediction errors to assess performance.

  • Test forecasts across multiple periods: Explore rolling evaluations and expanding windows to understand how forecasting performance changes over time.

  • Complete a practical forecasting project: Combine data preparation, model fitting, prediction, evaluation, and result saving into a reusable Python workflow.

A Practical Approach Designed for Beginners

Each chapter introduces essential concepts before moving into implementation. You'll find understandable explanations, complete Python examples, guided exercises, and expected results that help you verify your progress.

No advanced mathematics, statistics, or machine learning background is required. Basic familiarity with Python is helpful, but the forecasting tools and procedures are introduced progressively.

Whether you're an aspiring data analyst, Python learner, student, or professional interested in data-driven planning, this book provides a structured foundation for understanding and applying time series forecasting.

Move beyond simply generating predictions. Learn how to prepare the data, build the forecast, measure its errors, and understand what the results actually mean.

Format:
Paperback
Language:
English
ISBN:
B0HRD3MV35
ISBN13:
9798141932519
Release Date:
November 2026
Publisher:
Independently Published
Length:
232 Pages
Weight:
0.90 lbs.
Dimensions:
0.5″ × 7.0″ × 10.0″
ComputersComputers & TechnologyMathMathematicsScience & Math

More by Niels Heilmann

Customer Reviews

0 customer ratings | 0 reviews
Write a Review

There are currently no reviews. Be the first to review this work.