You work as a machine learning specialist for the sales department of a large web retailer that needs to gain insight into their sales patterns.
They need a way to use a visualization to show their sales data in near-real time so that they can quickly recognize higher-than-expected sales of specific products.
This will help your product operations quickly meet high demands.
Which option is a viable, efficient solution to your problem?
Click on the arrows to vote for the correct answer
A. B. C. D.Correct Answer: C.
Option A is incorrect.
Kinesis Data Streams cannot stream your data directly to S3
Also, running your own SageMaker Random Cut Forest model against your data is much less efficient than using the QuickSight ML Insights integrated Random Cut Forest capability.
Option B is incorrect.
While Kinesis Data Firehose can stream your data directly to S3, running your own SageMaker Random Cut Forest model against your data is much less efficient than using the QuickSight ML Insights integrated Random Cut Forest capability.
Option C is correct.
Streaming your data directly to S3 using Kinesis Data Firehose is very efficient.
Also, using QuickSight's integrated ML Insights Random Cut Forest capability requires far less development and coding effort than the other options.
Option D is incorrect.
Kinesis Data Analytics has a Random Cut Forest capability that you can use to detect your sales outliers.
However, you would still have to build your visualization in QuickSight.
The option of using ML Insights directly within QuickSight allows you to run your anomaly detection and visualize your data more quickly.
References:
Please see the Amazon QuickSight user guide titled Working with ML Insights (https://docs.aws.amazon.com/quicksight/latest/user/making-data-driven-decisions-with-ml-in-quicksight.html),
The Amazon QuickSight user guide titled Detecting Outliers with ML-Powered Anomaly Detection (https://docs.aws.amazon.com/quicksight/latest/user/anomaly-detection.html),
The AWS Machine Learning blog titled Visualizing Amazon SageMaker machine learning predictions with Amazon QuickSight (https://aws.amazon.com/blogs/machine-learning/making-machine-learning-predictions-in-amazon-quicksight-and-amazon-sagemaker/),
The Amazon Kinesis Data Analytics SQL reference titled RANDOM_CUT_FOREST (https://docs.aws.amazon.com/kinesisanalytics/latest/sqlref/sqlrf-random-cut-forest.html)
The best solution for the given scenario is option D: Use Kinesis Data Firehose to stream your data to a Kinesis Data Analytics application that runs a Random Cut Forest SagaMaker model on the data continuously, writing the output to S3, which is then used as source data to visualization in QuickSight.
Option A suggests using Kinesis Data Streams, which is a real-time data streaming service that allows data to be sent from a large number of sources to one or more destinations. However, Kinesis Data Streams requires custom application development to consume data and analyze it. Additionally, it is not a good fit for the scenario as it is costly and complex to set up and maintain.
Option B suggests using Kinesis Data Firehose to stream data to S3. Kinesis Data Firehose is a fully managed service that can capture, transform, and load streaming data into AWS data stores, data warehouses, and analytics tools. However, it does not provide any built-in analytics or data processing capabilities.
Option C suggests using QuickSight ML Insights, which is a machine learning-powered feature of Amazon QuickSight that helps discover insights and outliers in data using statistical models. However, it is not suitable for the given scenario as it is not designed for real-time data analysis.
Option D suggests using Kinesis Data Firehose to stream data to a Kinesis Data Analytics application that runs a Random Cut Forest SagaMaker model continuously. Kinesis Data Analytics is a fully managed service that can process and analyze streaming data using SQL, Java, or Scala. It can also integrate with machine learning frameworks like Amazon SageMaker to perform real-time analytics on streaming data. The output of the data analysis can be written to S3, which can then be used as a data source for visualization in QuickSight. This solution is efficient, scalable, and cost-effective, making it the best option for the given scenario.
In summary, option D is the best solution as it uses Kinesis Data Firehose for data streaming, Kinesis Data Analytics for real-time data analysis, S3 for data storage, and QuickSight for visualization.