Text Analytics and Keyword Analysis
Today we're going over Social Media Text Analytics for Business Intelligence. We will explain the several reasons why social managers analyze social text and will provide an example on how to perform a keyword search and analysis.
Social Media Text Analytics
Purpose of Social Media Text Analytics
Khan, Gohar- Social Media Analytics, 2018
Text analytics is used to examine and discover: sentiments, trends, intentions, stance, and concepts.
Sentiment analysis is used to determine how customers feel about a product or service. To perform this, you must compare the keywords in social media to lists of words that are positive or negative.
Trends mining uses large amounts of data to predict patterns and trends for a new product or service. This kind of mining uses machine learning and data mining.
Intention mining is used to analyze social media language and figure out the users' intention, like buying, selling, wishing, etc. Companies perform this so they can find new customers who are interested in their product.
Stance detection is a way to automatically detect if a piece of social media text is in favor, against, or neutral, or a specific topics.
Concept mining is used to extract ideas from large amounts of text that lives on web pages like Wikipedia, news transcripts, etc.
Twitter Keyword Analysis
For this analysis, I used a free-trial version of Sprout Social. I conducted a keyword search on Twitter, comparing the top 5 most used video editing softwares: Adobe Premiere, Avid Media Composer, Filmora, Final Cut Pro, and iMovie.
I selected the search to report from 08/01/21 to 08/31/21. At the top of my report was the keyword volume, by day. Looking at the layered keywords, there is one small spike in keyword mentions for "Filmora" and one large spike for "Adobe Premiere". To understand why there is a spike in conversation and keywords, I have to search the web for answers.
Digging into Google with the keyword "Filmora" and the date of the spike, around August 10th. I was able to find that Filmora had released an update that resulted in a lot of online talk, especially in Japan. You can see a sample of the Twitter talk below. Keep in mind, in the graph below, the volume numbers are different so the graph appears larger.
Looking into the large "Adobe" keyword spike for August 25, I found out there was a recent update and a Video Community Meetup on YouTube. In my research I discovered that these Adobe Video Community meet-ups feature product demos, conversations, and Q&A with Adobe product teams, and spotlights exciting projects from the community. Again, you can see below the social media talk that occurred:
Share of Volume
Onto the Share of Volume, we can see in this report that the largest volume of the 5 keywords was Adobe Premiere at 40%. iMove comes in second for share of volume at 28%. Even though they didn't have a spike in usage, iMovie is talked about more often generally than Filmora. Filmora trails right behind iMovie at 24% share of volume.
If we take a look at the Statistics by Keyword, we can dive deeper into the total volume of Tweets and the growth trend.
Now the simple keyword analysis is a great comparison to see what brands users are talking about most. This doesn't dive deep into sentiments, trends, intentions, etc. In order to explore those, more research and analytics data are needed. It seems that product releases or updates generally cause an uptick in conversation among users. This is assumed that the social chatter around updates are questions around the update and/bugs that may come along with them. I hope that this brief analysis helps you understand the very basics of text analytics!