I am glad you turned back in! Wait, didn't read Part 1? Go back and take a look, i'll wait.
After hearing about Text-Similiarity, today you will hear all (or something) about Sentiment analysis in a rather quick and dirty way.
What is it, and why is it being used/ needed?
Different articles on twitter, facebook, The Washington Post, etc are pretty subjective (Convince me of the opposite in the comment section). By subjectiveness i am referring to the Sentiment you receive when reading the article. A succesful flight to the moon will generally be more positive, since science has surprised us once again where as a beginning war in China will result in a lot of negative statements and hence be negative.
Determining the sentiment of an article is pretty straight forward. Let's take the following setence as an example
"Science has brought us positive light towards our future. We are happy to see how climat change wont effect us anymore as it used to do. Tremendous luck is involved."
Looking at the words, you cant directly spot "positive", "happy" and "luck(y)" which all belong to positive adjectives. Counting all adjectives in this sentence and calculating how many of them were positive, you will get a pretty good indicator what the sentiment is. The same accounts for sentences with a lot of negative vibes like "bad", "disastrous", etc etc..... Often times based on these values between -1 (negative) and 1 (positive) there can be done some conclusions i.e. trading stocks and cryptocurrencies based on twitter sentiment. (I won't guarantee success).
This one was pretty short but summarizd all you had to know in order to understand how Sentiments are being concluded. Today there is no TL;DR since this article is as short as an extra conclusion. My next article will focus on recommendation engines and categorization of for example customers and products.