I started this post in order to help anyone who is struggling with a similar problem. The initial work was done for a company that was doing this work every day.
Few employees had to go on amazon every day and get all the negative reviews details for some products and contact the reviewers.
This can be changed very easy to get all the negative reviews from youtube, facebook or any other website. For youtube, facebook or other website, sentiment analysis must be used to determine the sentiment of the comment.
Going back to amazon use case, I use a “List enumerator” to go over each review and get the title, content and the date. For list enumerator to work properly, you have to select 2 elements each from a different review. For instance, select the review content from review number 1 and review content from review number 2.
After this, any element that is inside list enumerator will resolve to the parent element, which means a “button click” activity or “get html element” will be on the current enumerated element. First time, they will be resolved against the first review, second time against the second review, etc.
Each extracted piece of data is saved in a datatable and at the end the data table is saved to a csv file.
The example can be extended to send an email, analyze the content or upload the data to some crm system.
Below is a gif recorded when running the above sample