Top sites allow us to see where is the conversation coming from, so drilling down a little bit more into those content sources really. And primarily, it's coming from Twitter, 94% of the data coming from Twitter. Some of the conversation coming from Reddit, some from Facebook. And then we get sites that account for a significantly smaller fraction of the conversation. As with the other sections on Crimson Hexagon, we are able to export this data. Now, as Crimson Hexagon's a certified partner for Twitter, there's a little bit more functionality built in if you're interested in primarily digging into the social media activity that's occurring on Twitter. So, if we click on the Twitter link under the metric section, top mentions of which individuals. If we expand this a little bit, we can see which Twitter accounts are counting for top mentions. If we look at hash tags, we can see which hash tags are counting for the bulk of the conversation. And then we also can look at top retweets. And so, we can see, looks like celebrities driving some of this. Taylor Swift with Mistaken song lyrics driving the most retweeted tweet that was over this particular time frame. And so, a little bit more detail as far as what we're going to be able to get from the Twitter activity. And we'll take a look at, and we can see here there is an Excel option for exporting this data. We'll take a look at, what does the raw data data look like, so the social media links we'll pull that out from the post list just to get a broader sample. But you'll see the level of detail that we're able to get. In our next module we'll talk about how we might go about analyzing some of the data that we extract. So we've looked at the volume and sentiment metrics. The next thing I want to walk you through are some of the content analysis tools that are built into the buzz monitor. And many of these tools are going to carry over to the social account monitor, as well as the opinion monitor. So, if we start looking at the specific language that's being used in the posts, that's going to give us a little bit more information than just how many posts are there, what's the volume or what's the sentiment of the posts? And so, this is the topic that Crimson Hexagon provides and you'll see that it's based only on 1,000 posts. If we want to add more posts, we can click on load more posts and that'll probably change things slightly though I would not expect it to change things all that much. And so, if we look at Starbucks coffee, that inner ring, okay, that's what a lot of the comments relate to and let's drill down further into that. Well, it relates to specific topics within Starbucks coffee. If we're looking at working at Starbucks, again you see jobs and openings, Starbucks Baristas, that accounts for another chunk of the conversation. So, we can get a sense for some of the conversation happening at Starbucks, but I think two of the other tools are going to be a little bit clearer. So let's start with the one that's probably the most common, is the Word Cloud Now, what the Word Cloud does is it's simply based on word count frequency. And so, the bigger the word in this diagram, the more frequently it's mentioned. Again, based on 1,000 posts, Crimson Hexagon will allow us to load more posts. So now it's based on 2,000 posts, and you can see it's relatively stable as we load more posts. And one thing that I want to show you is, this is giving us a sense for what's the most popular. If we want to look at exactly how frequently is this mentioned, let's click on the download and we'll download the Excel file. And let's open that up. Now, the Excel file is not going to download the Word Cloud for us. If you're interested in the image, you're going to want to download the image file itself. But what this is going to give us is the word frequency, so word count frequency. And that's the data on which these word clouds are based. And you can see for the particular time frame that we selected beginning in 2015 through 2016, and I'll zoom in so that we can see in a little bit more detail how frequently are specific words mentioned. So, let's just do a quick sort based on occurrence. And so, if we highlight and sort based on currents and I want to sort from largest to smallest. So what I'm going to do is, what are the most frequently mentioned words all the way down to what are the least frequently mentioned words in this sample. So, this is based on 3,000 posts. The highest frequency is using the reference to the Starbucks Twitter account, coffee, like, job, the Starbucks hash tag being mentioned frequently, barista being mentioned frequently. As we scroll down to the bottom, some words that are mentioned as infrequently as 12 times showing up in our Word Cloud. So that's the data that the Word Cloud is based on. Now if something that you wanted to look at is, how does a particular word change over time? One way to approach, or the use of that word change over time. One way to approach that would be to say, let's look at weekly or daily windows and pull the word count frequency for each of those windows to be able to compare them. Another way that we can do that, though, is if I wanted to apply a filter. So, I could apply a filter. So, I only want to look at mentions of the word, coffee. Now if I create that filter. Now if I apply this filter, what it's going to do is say okay, well let's restrict what we're looking at to only those posts that mention the word coffee. Now, that filter' s still populating. So, that would be the way in which we would be able to say, okay, well, let's look at volume for relating to only a subset of the posts. So I'm going to take that filter off, we'll go back to all of the data that we have to work with. So Word Clouds are one way of looking at the data. Another way that I like are the Word Clusters that Crimson Hexagon offers. Now, in contrast to the Word Cloud, what the Word Cluster does is allow you to see some of the connections that exist. And let's go back to the original version of the Word Clusters. And so, you see Starbucks dominates the conversation because recall that's what our search was based on, the word Starbucks. And so, what I'm going to do is I'm just going to click on this x and that's going to eliminate Starbucks from the Word Cluster, and reform the clusters based on that. And so, you can see, what are these word clusters based on? Well, let's click and drag this away. And so, one cluster based on coffee liking coffee today, and then if I look at another cluster, let's just drag that away to create some space. You can see it seems like it's related to the drinks. We see frappuccino, latte, pumpkin and spice being mentioned together. Another set of clusters appears to relate to jobs at Starbucks. And then it seems like we've got a decent amount of volume relating to the barista and ordering and through, perhaps, referencing the drive through. And so, we do get a little bit of flexibility here as far as how do we want the clusters formed. Let's include more than 45 words. Maybe we want to include 70 words, and let's see if that changes things. Little bit if we click on barista, drink, order, name. It's still the same general clusters though. The job cluster, the coffee cluster. It looks like we've our frappuccino, caramel, pumpkin spice, latte cluster, gift cards, looks like the word free, so maybe relating to sweepstakes. So, we get some flexibility around how these clusters are created. And this is a nice step in the direction of being able to understand, not just one word at a time being mentioned in these posts, kind of the bag of words analysis, but how the words are used together. Now, if we're interested in the raw data itself, we can click on the topic list. This is only going to give us a sample, but if we scroll down to the bottom, you can see this is only 100 posts. I'm going to download that sample of 100 posts just to show you what the post level data looks like in Crimson Hexagon. You can work with Crimson Hexagon if you need more than the sample that you're provided with. You can load more posts in here. You'll notice from our home screen that there's a limit to how many raw posts we're able to export, and that's part of the contracts negotiated with Crimson Hexagon.
