Yes using key words to come up with emotional impact score is not as crazy as you might think and yes it could be an indicator of near future trends of people but it isn't foolproof.

You need to manually set aside words and categorize based on emotional impact.

For example:

OK fine good great excellent

All these words could have the same meaning in a sentence but "excellent" goes much higher on the emotional score than "ok". Now imagine you have a list of English Language words that are all categorized and given a point system (" horrible" is -5, "OK" is +1, "excellent" is +5, etc). Now you have a program that automatically reads articles and analyzes the sentiment of the article based on points. More positive sentiment gives more positive points. You can take words that are in categories and give a sentiment of an article like "sad", " happy", "angry", "curious", etc, but that is likely less accurate (and lucrative) than just a positive or negative sentiment.

Yes using key words to come up with emotional impact score is not as crazy as you might think and yes it could be an indicator of near future trends of people but it isn't foolproof.

You need to manually set aside words and categorize based on emotional impact.

For example:

OK fine good great excellent

All these words could have the same meaning in a sentence but "excellent" goes much higher on the emotional score than "ok". Now imagine you have a list of English Language words that are all categorized and given a point system (" horrible" is -5, "OK" is +1, "excellent" is +5, etc). Now you have a program that automatically reads articles and analyzes the sentiment of the article based on points. More positive sentiment gives more positive points. You can take words that are in categories and give a sentiment like "sad", " happy", "angry", "curious", etc, but that is likely less accurate than just a positive or negative sentiment.