Doubt regarding streaming data analysis question

[Arghya Chakraborty]

Question: A leading online media company runs a popular sports news website. The solutions architect has been tasked to analyze each web visitor’s clickstream data on the website to populate user analytics, which gives insights about the sequence of pages and advertisements the visitor has clicked. The data will be processed in real-time which will then transform the page layout as the visitors click through the web portal to increase user engagement and consequently, increase the revenue for the company.

Which of the following options should the solutions architect implement to meet the above requirements?

CORRECT ANS: Push web clicks by session to Amazon Kinesis and analyzing behavior using Amazon Kinesis workers

INCORRECT ANS: Publish the web clicks to AWS IoT Analytics. Run custom analysis, SQL queries and apply machine learning to generate relevant reports regarding user behavior.

Reason given for incorrect answer: The following options are incorrect as SQS, EC2, EMR services do not have the capacity to analyze real-time streaming data

– This reason does not make sense. Could you please clarify why we can not use AWS IoT to collect & analyze the clickstream data ?

[ccatchings]

I agree that the explanation for the question is lacking and could better explain why using IoT Analytics is an incorrect answer here. However, this is not a good use case for AWS IoT in general as it talks about a website, not devices. AWS IoT Analytics uses “channels”, which devices uses to push raw data to an MQTT topic. All this solution needs is to push click events from a website to a real-time data stream. The answer involving pushing webclicks to Kinesis is the best one here (workers here are basically Kinesis Client Library consumers).

[Carlo-TutorialsDojo]

Hello Arghya,

Thanks for posting here.

Just like what ccatchings has said, AWS IoT is a service that you’d use for collecting data coming from IoT devices, such as sensors, wearables, etc. It’s not something you’d use for streaming a sequence of clicks on a web page. Also, in IoT, communication typically occurs over the MQTT protocol, whereas clickstream data on the web is usually transmitted using HTTP.

Let me know if this helps.

Regards,

Carlo @ Tutorials Dojo