What does this AWS Solutions Implementation do?

The Discovering Hot Topics Using Machine Learning solution helps you identify the most dominant topics associated with your products, policies, events, and brands. Implementing this solution helps you react quickly to new growth opportunities, address negative brand associations, and deliver higher levels of customer satisfaction.

The solution uses machine learning algorithms to automate digital asset (text and image) ingestion and perform near real-time inferences using topic modeling, sentiment analysis, entity and key phrase detection, and detection of unsafe images. The solution then visualizes these large-scale customer analyses using an Amazon QuickSight dashboard. This guide provides step-by-step instructions to build a dashboard that provides you with the context and insights necessary to identify trends that help or harm your brand.

The solution performs the following key features:

  • Performs topic modeling to detect dominant topics: identifies the terms that collectively form a topic from within customer feedback
  • Identifies the sentiment of what customers are saying: uses contextual semantic search to understand the nature of online discussions
  • Determines if images associated with your brand contain unsafe content: detects unsafe and negative imagery in content
  • Helps customers identify insights in near real-time: you can use a visualization dashboard to better understand context, threats, and opportunities almost instantly

The solution deploys an AWS CloudFormation template that supports Twitter as the default data source, but the solution can be customized to aggregate other social media platforms and internal enterprise systems.  

After you deploy the solution, use the included Amazon QuickSight dashboard to visualize the solution's machine learning inferences. The image to the right is an example visualization dashboard featuring a dominant topic list, donut charts, weekly and monthly trend graphs, a word cloud, a tweet table, and a heat map.

Discovering Hot Topics using Machine Learning | Sample Dashboard
Sample Amazon QuickSight dashboard
 Click to enlarge

AWS Solutions Implementation overview

The diagram below presents the serverless architecture you can automatically deploy using the solution's implementation guide and accompanying AWS CloudFormation template.

Discovering Hot Topics using Machine Learning | Architecture Diagram
 Click to enlarge

Discovering Hot Topics using Machine Learning solution architecture

The AWS CloudFormation template automatically deploys AWS Lambda functions, Amazon Simple Storage Service (Amazon S3) buckets, Amazon Kinesis Data Streams, Amazon Kinesis Data Firehose, AWS Step Functions workflows, and AWS Glue tables in your account.  

The architecture of the solution includes the following key components:

Ingestion – Social media feed ingestion and management using Lambda functions, Kinesis Data Streams, and Amazon DynamoDB.

Workflow – A Step Functions workflow that orchestrates Amazon Translate, Amazon Comprehend, and Amazon Rekognition.

Inference – AWS machine learning capabilities through Amazon Translate, Amazon Comprehend, and Amazon Rekognition.

Application Integration – Event-driven architecture using Amazon EventBridge.

Storage and Visualization – A combination of Kinesis Data Firehose, Amazon S3 buckets, AWS Glue, Amazon Athena, and Amazon QuickSight.

These components are built using the AWS Well-Architected Framework, and the AWS Well-Architected Pillars of Operational Excellence, Security, Reliability, Performance Efficiency, and Cost Optimization—ensuring secure, high-performing, resilient, and efficient infrastructure.

Discovering Hot Topics using Machine Learning

Version 1.3.0
Last updated: 11/2020
Author: AWS

Estimated deployment time: 10 min

Use the button below to subscribe to solution updates.

Note: To subscribe to RSS updates, you must have an RSS plug-in enabled for the browser you are using.  

Did this Solutions Implementation help you?
Provide feedback 


Near real-time analytics

Ingest streaming data containing text and images, then analyze them in near real-time.

Multi-lingual data ingestion

Use Amazon Translate to ingest data in multiple languages.

Machine learning

Use Amazon Translate, Amazon Comprehend, and Amazon Rekognition to perform topic modeling, sentiment analysis, entity detection, key phrase detection, and analyze images for embedded text and unsafe content. The text detected in images is also used for topic modeling, sentiment analysis, entity detection, and key phrase detection.

Secure one-click deployment

Provide a secure one-click deployment using an AWS CloudFormation template developed with the AWS Well-Architected Framework methodologies.
Solving with AWS Solutions: Discovering Hot Topics Using Machine Learning
Build icon
Deploy a Solution yourself

Browse our library of AWS Solutions Implementations to get answers to common architectural problems.

Learn more 
Find an APN partner
Find an APN Partner

Find AWS certified consulting and technology partners to help you get started.

Learn more 
Explore icon
Explore Solutions Consulting Offers

Browse our portfolio of Consulting Offers to get AWS-vetted help with solution deployment.

Learn more