Let's bring things together in this last video. Finance, data, technology, as we said at the outset are intimately related. Really since almost the beginning of finance, data has played a central role and as technology has progressed, it has creeped into finance, at least the practice of finance, not to mention the academic study almost immediately. There's this nice combination among the three areas that has led us to where we are today in which concepts such as machine learning and AI are playing a central role. What we've done in this set of videos is we've really talked about a variety of things, all of which are related in a linear way. Problem-solving starts with the scientific method clearly articulating a question, being very precise, hypothesizing answers to the question, understanding the implications of those answers for what you might see in the data, and then taking those hypotheses to the data by way of the data science workflow. Beginning with acquisition and verification of the data, followed by preparation of the data, getting it ready for analysis, performing exploratory data analysis on it, and then eventually modeling. We showed all of this in the context of an application just as an illustrative vehicle, specifically corporate credit risk. We talked about what corporate credit risk was, how important it is, not just for firms, but for a variety of different stakeholders in firms. We walked through the basics of assessing credit risk in a firm by looking at its financial statements, by looking at key performance indicators, then we transitioned into talking about credit ratings, recognizing the credit ratings are getting us towards a quantification of credit risk built on the analysis that we started off with in terms of financial statement analysis. Then we actually decided to model credit ratings, all be it at a somewhat course level of looking at speculative-grade versus investment grade, and we implemented a machine learning process starting with a logit model and exploring a variety of other models, including k-nearest neighbors. We avoided the technical details because this is neither the time nor place for that, we just want a general overview introduction to some of these concepts. There's tons of information elsewhere on the gory details for those that are interested. But one of the lessons that came out of this that isn't often emphasized elsewhere is the importance of data versus models. If you have to spend resources, spend it on the data, not on fancier models because models are limited by what goes into them, period. The other thing we learned is it's just as important to understand where you're making mistakes. Arguably more important than where you're being successful because that's where improvement is going to come in terms of the modeling process. Now, through all of this, you might be wondering, what about AI? We haven't talked about AI, artificial intelligence. I like to think of AI as really a superset that encompasses machine learning, but other stuff as well. To keep things digestible in this video series under a reasonable time cap, let me just speak a little bit off the cuff here. There's many different roles that AI can play in finance. One that I think is worth emphasizing is in overcoming or mitigating some of the behavioral biases that are inherent in human decision-making. Look at some recent Nobel Prize winners, whether it's Danny Kahneman or Richard Thaler, behavioral economics, behavioral finance more broadly is so important because it has shone a very bright light on the mistakes that individual investors, humans more broadly make when it comes to economic decisions or financial decisions. We make mistakes all the time. What I think AI can do among many other things, is it can really help discipline decision-making, it can help us avoid making some commonsense mistakes when it comes to investing decisions or other finance-related decisions. There's enormous scope and we're already seeing some of this in Robo-investing advice. But I'm not going out on any limb in saying that that is only going to grow in the role for AI in helping people make better financial decisions is only going to become more and more important as time progresses, as we get more data, and computing power ramps up. Let me just stop talking. Thank you for your time. Say I hope you find the videos useful if not enjoyable.
