AI-Powered Time Series Forecasting with Python

Course details

For any business, gaining and understanding insights into future trends, customer demands, or market conditions is an important factor in success. And with the wide availability of machine learning and artificial intelligence tools, thousands of businesses are able to enhance their operations through time series forecasting. In this course, Tobias Zwingmann introduces you to time series forecasting using Python and AI and shows how you can apply them to your business. Learn how to translate forecasting workflows from static, classroom problems to dynamic, real-time use cases. Plus, find out about the tools and approaches you can apply to other AI and machine learning tasks.

Learning objectives
Learn the differences between batch and online prediction, and what each is best suited for.
Discover the various types of features that go into a real-time model—online, batch, session-based, and more—including how to calculate and store each given our example.
Learn about real-time vs. near real-time prediction and the differences between both paradigms and when to use each.
Learn the importance of model monitoring and how to effectively use it to keep models fresh.

Concepts

0. Introduction

01 - Introduction
02 - What you should know
03 - GitHub Codespaces
04 - Model walkthrough

1. Batch Systems

05 - What are batch features
06 - Getting started with batch features in Codespaces
07 - Building a store for batch features
08 - Training our model to predict
09 - Making predictions with our model
10 - Advantages and disadvantages of batch forecasting
11 - Challenge - Feature X
12 - Solution - Feature X

2. Near Real-Time Systems

13 - What are near real-time systems
14 - Requirements for near real-time forecasting systems
15 - Recalculating features
16 - Frequency considerations
17 - Online prediction
18 - End-to-end example
19 - Advantages and disadvantages of near real-time
20 - Challenge - Feature Y
21 - Solution - Feature Y

3. Real-Time Systems

22 - What are real-time forecasting systems
23 - Requirements of real-time forecasting systems
24 - Streaming datasets
25 - Online features
26 - Online prediction
27 - End-to-end example
28 - Real-time forecasting and latency considerations
29 - Advantages and disadvantages of real-time forecasting
30 - Challenge - Feature Z
31 - Solution - Feature Z

4. Evaluating Time Series Forecasting Systems

32 - Evaluating forecasting models
33 - Best practices for retraining time series models

Conclusion

34 - Next steps for AI forecasting

AI-Powered Time Series Forecasting with Python

Authors

Tobias Zwingmann