Now that we talked about the theoretical foundation for brand personality and also talked about a machine learning model that can estimate brand personality from Twitter data, let's jump into the Social Media Macroscope and look at some brand personalities. But before we do so, please know that we will be using a different tool, not the SMILE tool in the Social Media Macroscope, but the BAE tool, the brand analytics environment tool. I also want to say before we jump in, that more evidence is needed to validate both the machine learning model for brand personality prediction, as well as this potential personality congruence phenomenon where people seem to follow brands that they're close in personality to than brands that they're not. This is all just stemming from one study that we did. Many more studies are necessary. But it's interesting enough where, I believe, we can try to use some of these methods to see if that can affect how we find influencers that might affect customer satisfaction towards our company brand, products, and services. With that all said, let's jump into the Social Media Macroscope. Here's the dashboard for the Social Media Macroscope. You've been here before because you've launched SMILE, and so you've either click on this box icon next to SMILE or you've clicked directly on SMILE and gone into the launch app button. But we are going to use the brand analytics environment, which was built for that study that I had mentioned in the lesson previously, that involved calculating people's human personality using IBMs tool, and then testing the congruence of personality with brand Twitter accounts that these people followed versus brand Twitter concept they did not follow. Let's start brand analytics environment. Let's click on this box either to the right or you can click on Brand Analytics Environment here and click on "Launch Tool", and just shortly the tool will launch itself. Now there's currently two personality models in BAE. One is the human personality model that IBM created. We will not be using that model, but rather the model that my team created. Right here, again, we start with citations. This is actually the paper, if you'd like to reference it with regards to our previous study. We also asked for a citation because we are trying to present these tools for free, and so we need support and proof of applicability to the real-world as well as to research. If you are doing research, will you please consider citing us. Here, we authorize our Twitter account. Let's click on "Authorize App", let's grab the pin, spring the pin and this is just like SMILE. But then for this IBM stuff, because we're not going to be using IBMs personality model as well as you need model API keys any ways that you do not have, let's click on "Skip" here. Then what we'll do is we'll choose the Pamuksuz, Yun, and Humphreys brand personality model. Let's compare a influencer on Twitter. We'll choose Kim Kardashian. As you type in letters, it'll just autofill from Twitter. Then let's choose Apple. Interesting thing about Apple's main account is they don't tweet from that account, so it has zero tweets. Maybe what we can choose is we can choose something like Apple TV or Apple Arcade. Then what you'll do here is you'll click on "Start Analysis". When you click on Start Analysis, it will ask you for your email address because this computation takes quite a bit of time. You'll put in your email address, and then you'll click on "Submit". Then once you click on Submit, you'll get an email once it's done. Now, I've already ran this analysis so I'll just go to my history. But just FYI, it will take quite some time for you to get an email to allow this to fully run. The way I'll access my previous results is to go to history up here in this top left-hand corner. I'll click on History, and you'll see here that I have the Kim Kardashian Twitter account official Pamuksuz results. Here are the values for Kim Kardashian in sophistication, excitement, sincerity, competence, and ruggedness. Again, these values by themselves don't mean that much because really what we need to do is compare them to other accounts. Let's say we look at Apple Arcade here, and we see values also for sophistication, excitement, sincerity, competence, and ruggedness. Remember, this is now taking tweets that Apple Arcade has presented on their Twitter account and running that through our brand personality model. How we can test congruence is we'll choose the Pamuksuz, Yun, and Humphreys brand personality model, and then we'll choose a few different accounts. We'll choose KKHOfficial. Let's choose Apple Arcade. Let's add another one even, Apple TV, and we click on "Bulk Comparison" here. What we'll see here is that we've got a matrix that compares the personality congruence using cosine similarity of various accounts. If you compare, for example, Apple TV to Apple TV, it's a perfect, 100 percent, 1.0 comparison, and that makes sense. Apple Arcade to Apple, Arcade 1.0. Kim Kardashian to Kim Kardashian, 1.0. Now let's look at Kim Kardashian to Apple TV. It's a 94 percent or 0.94 similarity. Let's also look at Kim Kardashian to Apple Arcade, it's 81 percent, 0.81 percent similarity. Again, drawing from previous research, there might be some validity to Kim Kardashian potentially participating with Apple TV because the way that they tweet is very similar in brand personality, that there might be some potential interaction that could occur here. Of course, it could not just be Kim Kardashian, but you could choose anybody. Maybe you find somebody using network analysis for your brand on Twitter. It might be somebody that might not necessarily be a big celebrity but somehow they've got very high influence with regards to network structure on Twitter. Then you run their brand personality and see if it matches the brand personality of your brand. If it's a high number like this, 94 percent, then you may consider contacting them and partnering with them to market your brand because, again, this is all about potentially trying to move the needle with regards to customer satisfaction. We know that influencers have a pretty large effect on customer satisfaction towards a brand, especially on social media.
