Chapter 7. Building Predictive Analytics for AWS IoT

Machine learning is one of the potential technologies that has multiple benefits when applied to our business cases. In this chapter, you will learn various Amazon AWS services related to machine learning. Some demos will be explored using machine learning services from Amazon.

The following is a list of topics covered in this chapter:

• Introducing AWS Machine Learning services
• Making your sensor speak
• Integrating Amazon Echo into your IoT projects
• Making image and video analysis
• Making predictive analytics for IoT data

Introducing AWS Machine Learning services

Machine learning is one of the subjects in computer science that describes how to teach computers to obtain the ability to learn without hardcoding in a program. More specifically, machine learning enables computers to learn from experiences such as data, information, and training results from experts.

All AWS Machine Learning services can be found at https://aws.amazon.com/machine-learning/. At the time of writing this book, AWS Machine Learning services are available in any region. Make sure that you change AWS region in order to enable AWS Machine Learning services:

In this chapter, we will explore some AWS Machine Learning services that are related to IoT topics.

Making your sensor speak

Suppose that you have deployed sensors at various locations. Those sensors perform monitoring to detect a specific purpose. When a sensor finds a thing that its looking for, a sensor device will send an alert to a command center. Some people probably stand by on that location. When an alert comes, the system can make a sound through a speaker. 

Message-to-sound or text-to-speech is a field of speech technology that leverages machine learning. A system can convert text or message to human speech. In general, the scenario can be described in the following figure:

From the preceding figure, IoT device can perform sensing and send data to AWS. IoT device can send an alert message to AWS to inform to the command center. AWS can convert text message to sound message. Once a command center receives a sound message, it performs to generate sound through a speaker.

Amazon AWS provides a service that applies speech technology, named Amazon Polly. You will learn how to work with...

Integrating Amazon Echo into your IoT projects

Amazon Echo is one of the Amazon products that uses speech technology. Amazon developed Artificial Intelligence (AI) to enable interaction with machines through speech. You can buy Amazon Echo on the Amazon website at https://www.amazon.com/dp/B06XCM9LJ4/. Unfortunately, Amazon Echo probably does not support for your country. A list of supported countries includes their features can be found on this website https://www.amazon.com/gp/help/customer/display.html?nodeId=202207000.

Amazon also provides a lite version from Amazon Echo known as Amazon Echo Dot. You can buy it on https://www.amazon.com/gp/product/B01DFKC2SO/. I bought this device from Amazon Germany. You can see my Amazon Echo Dot in the following image:

In this section, we will build a project to integrate between Amazon Echo and IoT board. In general, we develop a program based on a scenario in the following figure. We want to turn on/off the LED by giving a voice command. For instance...

Making image and video analysis

In this section, we learn how to analyze image and video. Machine learning can perform pattern recognition for a picture or a video. The result of machine learning analysis can be object detection or image analysis.

For instance, I have a still image; this image is analyzed using machine learning. It shows some identified objects, such as human, brick, clothing, and coat. You can see it in the following image. This process is done using Amazon Rekognition: 

In this section, we focus on how to use Amazon Rekognition to analyze image and video. Setting and configuring will be performed in the next section.

Introducing Amazon Rekognition

Amazon Rekognition is one of the Amazon services that is part of Amazon Machine Learning. You just register and activate your AWS account on this services. Amazon Rekognition can perform object and scene detection, image moderation, facial analysis, celebrity recognition, face comparison, text in image analysis, and video analysis...

Make predictive analytics for IoT data

Predictive analytics is a method to make prediction for an unknown event. In the context of IoT, we can develop predictive analytics to make a decision-based streaming sensor data. This is a part of machine learning study. In general, we can make predictive analytics using a diagram that is shown as follows:

Defining business problems is the first step to develop predictive analytics. Some problems probably need experts to make clear those problems. For instance, economics, biology, and volcanology. 

We also should prepare data in order to develop a model. This data should have high impact factors on the model.  When we develop a model, we also perform some steps such as defining targets, extracting derived features from data, fitting the model, and evaluating the model. In a real project, we probably make some iterations to ensure the corrected model.

After we developed the model, we can deploy the model into our system. This could be deployed in web...

Build a simple predictive analytics for your IoT project

In this section, we will develop a simple predictive analytics for IoT. Our project can be described in the following figure. We will perform sensing to acquire temperature and humidity from the sensor devices. Then, we send this sensor data to AWS Machine Learning to get a decision on whether the system will perform watering:

The project will focus on developing a machine learning model. Assume that we have temperature and humidity data from the sensor. We do not implement watering system, but we will make a decision system to trigger watering.

Next, we will implement the project with the following steps which are explained in detail in the upcoming sub-sections:

1. Defining a machine learning model
2. Preparing data
3. Building a machine learning model
4. Evaluating and testing a model

Defining a machine learning model

We will create a simple model for our project. We have two inputs—temperature and humidity. We also have historical data about sensor...

Summary

You learned several AWS Machine Learning services, such as Amazon Polly, Amazon Alexa, and Amazon Rekognition. Then, you also learned to build predictive analytics for IoT data.

In the next chapter, you will learn about AWS IoT security.