Tweeter Temp Check #3 : Result Overview
RESULT OVERVIEW
BY ASHUR BAROUTTA
We're finally ready to start looking at our data! The following is a portion of the 100 tweets we've pulled.
While the analysis column helps show the sentiment rank, we can intuit similarly by looking at polarity/subjectivity scores. So now we've gotten to a place where we can start seeing sentiment rank across N number of tweets, obviously, in terms of practical business applications, printing out a long list of tweets with the words negative, positive, or neutral next to them doesn't really drive home the value in using sentiment analysis. It is recommended to always keep readability in mind, and for that reason we've taken the 100 tweets and plotted their results to easily allow for quick and efficient analysis.
From the plot we can quickly identify the user's posts as being overall, mostly positive!
While this might seem simple and limited in scope, there are many ways to utilize these methods for meaningful insight. For example, while here we only pulled 100 tweets, imagine a company wanting to see how receptive the public is to their marketing efforts. They could pull all tweets containing a specific hashtag relevant to their campaign and then run sentiment analysis on them to try and identify how well received it is. Or think of how companies track feedback, they could compile all their reviews and run a sentiment analysis on the customer submitted feedback, allowing for quick overview of general sentiment for the company's public facing employees!
This type of project provides experience and confidence for us to apply this to many different use cases. We will continue exploring sentiment analysis outside of future projects as long as we have API access.
Love the real world extrapolation, really drives it home.
ReplyDeleteThanks!
DeleteI wanted to make sure people had an idea of where this could be valuable as opposed to just some neat computer magic.
I am glad it is well received.