Return to page


Time Series Forecasting Best Practices


By Jo-Fai Chow | minute read | October 15, 2021

Blog decorative banner image

Earlier this year, my colleague Vishal Sharma  gave a talk  about time series forecasting best practices. The talk was well-received so we decided to turn it into a blog post. Below are some of the highlights from his talk. You can also follow the two software demos and try it yourself using our H2O AI Cloud .

(Note : The video links with timestamps may not work correctly with your browser. If so, right-click on the links and open them in a new window/tab.)

Overview of Time Series Forecasting

Vishal first discussed some use cases in different domains and explained what makes time series forecasting a special case of machine learning.

Time Series Forecast Use Cases [Video @ 0:53]
www.h2o.ai2021/10/ts_001.png www.h2o.ai2021/10/ts_001.png
Time Series Forecast is a Special Case of Machine Learning [Video @ 1:50]
www.h2o.ai2021/10/ts_002.png www.h2o.ai2021/10/ts_002.png


Time Series Pipeline

A typical time series pipeline is shown below. Vishal focused on preprocessing and forecasting components of the pipeline in this talk.

A Typical Time Series Pipeline [Video @ 3:05]
www.h2o.ai2021/10/ts_003.png www.h2o.ai2021/10/ts_003.png


Time Series Preprocessing

Some time series forecasting methods require data preprocessing. Here are some common techniques:

Common Time Series Preprocessing Techniques [Video @ 4:06]
www.h2o.ai2021/10/ts_004.png www.h2o.ai2021/10/ts_004.png
Features and Patterns for Forecasting [Video @ 5:13]
www.h2o.ai2021/10/ts_005.png www.h2o.ai2021/10/ts_005.png


Commonly Used Time Series Models

Vishal also discussed some of the commonly used time series models like AutoRegressive Integrated Moving Average (ARIMA) and Exponential Smoothing.

Time Series Model – ARIMA [Video @ 7:16]
www.h2o.ai2021/10/ts_006.png www.h2o.ai2021/10/ts_006.png
Time Series Model – Exponential Smoothing [Video @ 10:32]
www.h2o.ai2021/10/ts_008.png www.h2o.ai2021/10/ts_008.png
www.h2o.ai2021/10/ts_009.png www.h2o.ai2021/10/ts_009.png
Time Series Model – Others [Video @ 12:50]
www.h2o.ai2021/10/ts_010.png www.h2o.ai2021/10/ts_010.png


Challenges and Practical Considerations

Here is one of the key takeaways   – time series forecasting challenges and practical considerations:

Time Series Model – Challenges [Video @ 14:18]
www.h2o.ai2021/10/ts_011.png www.h2o.ai2021/10/ts_011.png
Time Series Model – Practical Considerations [Video @ 15:23]
www.h2o.ai2021/10/ts_012.png www.h2o.ai2021/10/ts_012.png


Time Series Forecasting with H2O

Vishal then explained the inner workings of time series forecasting with our platform in great detail.

Automatic Machine Learning with H2O [Video @ 20:56]
www.h2o.ai2021/10/ts_013.png www.h2o.ai2021/10/ts_013.png
Time Series Modeling Approach and Feature Engineering [Video @ 21:57]
www.h2o.ai2021/10/ts_014.png www.h2o.ai2021/10/ts_014.png
Time Series Feature Transformers
www.h2o.ai2021/10/ts_034.png www.h2o.ai2021/10/ts_034.png
Time Series Feature Engineering – Lag Interactions [Video @ 24:09]
www.h2o.ai2021/10/ts_017.png www.h2o.ai2021/10/ts_017.png
Making Use of the Top-Performing Algorithms from AI Community [Video @ 25:37]
www.h2o.ai2021/10/ts_020.png www.h2o.ai2021/10/ts_020.png
Model Validation [Video @ 26:32]
www.h2o.ai2021/10/ts_022.png www.h2o.ai2021/10/ts_022.png
Predictions with Refit [Video @ 29:08]
www.h2o.ai2021/10/ts_026.png www.h2o.ai2021/10/ts_026.png


First Demo – TSA Checkpoint Travel Numbers

The first demo was about forecasting the number of passengers. You can find the data from Transportation Security Administration (TSA) here .

TSA Passengers Demo [Video @ 29:44]
www.h2o.ai2021/10/ts_028.png www.h2o.ai2021/10/ts_028.png
Time Series Model Explanation [Video @ 40:31]
www.h2o.ai2021/10/ts_030.png www.h2o.ai2021/10/ts_030.png


Second Demo – Demand Sensing Wave App

The second demo was a demand sensing Wave  application. You can find more examples from our AI Cloud  app store.

Demand Sensing Demo [Video @ 42:04]
www.h2o.ai2021/10/ts_027.png www.h2o.ai2021/10/ts_027.png
www.h2o.ai2021/10/ts_031.png www.h2o.ai2021/10/ts_031.png
www.h2o.ai2021/10/ts_033.png www.h2o.ai2021/10/ts_033.png



In short, time series models require data analysis, preprocessing, and hyperparameters optimization. Automatic machine learning (AutoML ) is a practical choice for time series forecasting as it can handle multiple constraints. For some use cases, it is useful to incorporate COVID data in model updates for better predictive power.

Key Takeaways

Time series AutoML with our AI Cloud platform  can:

  • capture multiple temporal behaviours
  • carry out extensive experiments with linear/non-linear/hybrid models
  • optimize hyperparameters while considering practical and real-world constraints

How to Get Started

Request a demo today. 


Jo-Fai Chow

Jo-fai (or Joe) has multiple roles (data scientist / evangelist / community manager) at Since joining the company in 2016, Joe has delivered H2O talks/workshops in 40+ cities around Europe, US, and Asia. Nowadays, he is best known as the H2O #360Selfie guy. He is also the co-organiser of H2O's EMEA meetup groups including London Artificial Intelligence & Deep Learning - one of the biggest data science communities in the world with more than 11,000 members.