Today, I’m here to show you that power of Vantage. Okay, what is Vantage? Vantage is an enterprise-grade software, which brings together the data warehouse, the data lakes, and analytics together. It’s available on the public cloud, the private cloud, as well as on premises, and in a hybrid environment. Vantage is built on world-class Teradata database technology, brings together multiple genres of analytics together, such as your machine learning, statistics, graph. And here’s the deal: All of these different analytics’ functions and all of these data that are available across multiple sources can be executed in SQL, in Python, in R; as well as, analytic visualizations can be delivered through components like Vantage Analyst and AppCenter. The best part of Vantage is that it is available across a wide constituency of people in the organization who want to be data driven: your data scientists, your business analysts, the lines of the business managers. Anybody who wants to be data-driven can use Vantage. And all of these insights can be operationalized in the platform. Not only are you able to deliver these insights on all of the data across these different analytics techniques, but they can also be operationalized in the same platform. This is what Vantage is. Okay, but now that I’ve told you what Vantage is, let’s actually take a specific use case and see what we can do with Vantage. So, here’s an example. Let me set the context for you a little bit. Here is an example of a retailer who wants to understand customer churn. And the way they’re defining customer churn is that there are members in the organization who cancelled their membership. Understandably, the retailer wants to know why people are cancelling their membership. Because if they can know why, and if they can predict who is likely to cancel their membership, they can take certain bits of corrective action. So, here’s an example of a path analysis application. This is a visualization. It’s a Sankey visualization. And there are different sets of activities which contribute towards membership cancelation. So, here’s one example, right? So, it seems like in this case, there are a lot of people who seem to making a complaint call, before they cancel their membership. So, what does this tell the retailer? It tells the retailer that clearly there seems to everybody something going on with the complaints. So, they look at this path and they say, “Okay, let’s figure out what’s happening with the complaints.” So, the next thing they do is they go and they want to understand the nature of these complaints. So, what they go into is a sentiment analytics application. They look at it and say, “Look, it seems like a lot of people seem to be confused with our policies. They seem to be sick. They seem to think that a lot of our policies, a lot of our—what we do for them is silly and they’re confused. They’re tired. It’s inconsistent, it’s expensive, it’s inappropriate. There’s a lot of confusion. And, to that point, the word “confused” comes up big in the word cloud. Now, what’s the point? Now, they’re able to see a path.
My name is Sri Raghavan. I have been in the analytics profession for over twenty years and have worked with hundreds of companies to deliver analytics’ implementations across many verticals and use cases. In the last couple of years I’ve been traveling the world to train thousands of customers on Teradata’s Vantage analytics platform. Today, I’m here to show you that power of Vantage. Okay, what is Vantage? Vantage is an enterprise-grade software, which brings together the data warehouse, the data lakes, and analytics together. It’s available on the public cloud, the private cloud, as well as on premises, and in a hybrid environment. Vantage is built on world-class Teradata database technology, brings together multiple genres of analytics together, such as your machine learning, statistics, graph. And here’s the deal: All of these different analytics’ functions and all of these data that are available across multiple sources can be executed in SQL, in Python, in R; as well as, analytic visualizations can be delivered through components like Vantage Analyst and AppCenter. The best part of Vantage is that it is available across a wide constituency of people in the organization who want to be data driven: your data scientists, your business analysts, the lines of the business managers. Anybody who wants to be data-driven can use Vantage. And all of these insights can be operationalized in the platform. Not only are you able to deliver these insights on all of the data across these different analytics techniques, but they can also be operationalized in the same platform. This is what Vantage is. Okay, but now that I’ve told you what Vantage is, let’s actually take a specific use case and see what we can do with Vantage. So, here’s an example. Let me set the context for you a little bit. Here is an example of a retailer who wants to understand customer churn. And the way they’re defining customer churn is that there are members in the organization who cancelled their membership. Understandably, the retailer wants to know why people are cancelling their membership. Because if they can know why, and if they can predict who is likely to cancel their membership, they can take certain bits of corrective action. So, here’s an example of a path analysis application. This is a visualization. It’s a Sankey visualization. And there are different sets of activities which contribute towards membership cancelation. So, here’s one example, right? So, it seems like in this case, there are a lot of people who seem to making a complaint call, before they cancel their membership. So, what does this tell the retailer? It tells the retailer that clearly there seems to everybody something going on with the complaints. So, they look at this path and they say, “Okay, let’s figure out what’s happening with the complaints.” So, the next thing they do is they go and they want to understand the nature of these complaints. So, what they go into is a sentiment analytics application. They look at it and say, “Look, it seems like a lot of people seem to be confused with our policies. They seem to be sick. They seem to think that a lot of our policies, a lot of our—what we do for them is silly and they’re confused. They’re tired. It’s inconsistent, it’s expensive, it’s inappropriate. There’s a lot of confusion. And, to that point, the word “confused” comes up big in the word cloud. Now, what’s the point? Now, they’re able to see a path. You saw that in the path. They see sentiments that are being expressed that are very negative, and then they say, “Look, I seem to understand these things are happening, but what I still don’t know is how can I predict that somebody is actually going to not be a customer any longer?” So, here, what they do is they want to model it. They want to—you can model customer behavior in many different ways. So, here, for instance, is an example of a machine learning model. Forget the guts of the machine learning model for a quick second. Well, what they want to understand is what’s the risk that a customer is going to churn, based on all the different activities that have already taken place from the path analysis, from the sentiments that are being extracted [ph?]. Let’s actually try to model behavior. The reason why you do that on these screens where you can actually pick and choose the different models that you have—so, here, for instance, I pick one type of machine learning model, but there are many other different types of models that are available. You can actually run these models and then you can see how well the model has performed, right? Because whenever you have created this sort of predictive model, you want to see, “Does my predictive model make sense?” And, here, it looks like your model has got a high rate of precision 92 percent of the time. And that’s a fantastic precision rate. What is my point here? My point here is that we used Vantage, native applications in Vantage, without any coding whatsoever, pulled it out from screens to be able to understand how customers behave in terms of path analysis, what they are talking about you in terms of sentiment extraction, and how you can predict the likelihood that they’re not going to be a churner or are going to be a churner with this machine learning model on many different kinds of data that are available. But before we get into some of the differentiators, let me show you one more thing, right? Sometimes analysts say, “Well, look, all these screens are fantastic, but I’d actually like to see the code.” No problem. So, here is an example of how you would run the same retail churn example that I said with membership purchase and membership cancellation using code. So, here’s an example of SQL code, where I’ve selected function nPath, the path analysis, directly from Vantage. And then I used the same sentiment extractor function also from Vantage. I just simply call it directly from Vantage. I run sentiments to get the positive versus a negative on each one of the sentiments that have been expressed and put it in a word cloud. And then, here is an example of another kind of model. I showed you DecisionForest. But, here, for instance, is a model called XG Boost, where I run it directly to predict the likelihood that someone has churned. So, here is the prediction. This customer, 194020, is predicted to churn, right? And this is how you operationalize. Now, what’s the advantage of doing it all in Vantage. Number one, when we do it across path analysis and sentiment extraction modeling, different kinds of data are brought into the picture: your structured data, your semi-structured data in terms of weblogs and online activity; and then your completely unstructured data. All of those things are brought directly in Vantage. One fell swoop in one platform. Number one. Number two, did you see all these different analytics techniques and screens that I showed you? These are machine learning models. These are sentiment extractors, these are path analytic applications. If you go outside of Vantage, you have to have three separate solutions to bring them with three different screens and three different user interfaces meant for three different personas in three different solutions that you have to pay for. None of that in Vantage. All of those things are brought into one platform in Vantage. And that’s a huge differentiator. All of these activities from insights delivery to operationalize on all your data across multiple personas happens in Vantage and that’s a core differentiator.