Greetings, fellow golfers out there, and a warm welcome to this lightning talk on craft your game with go. As I was prepping up for this presentation, I was looking for that aha moment or a punchline to start off with. I have kind of decided that I'll reserve it for the next slide. A little bit of context sitting here. How many of you have watched money Bahal the million dollar arm? How many of you still recollect this classic saber metric scene from Moneyball? A fun fact. These movies cumulatively cost 140 million at the box office and are based on sports analytics. What's the big deal about sports analytics? Largely the measure of a professional sports team's success has traditionally been pretty straightforward. The team either wins or loses. Hey, there's a lot more to it, Albert. What happens on the playing field, a lot dependent on what happens outside it, what decisions are taken outside it. I'm going to do a bit of a role play here, case study and look at two specific cases in the next couple of slides. You know, what is the problem at hand? What is the analytics problem at hand? Rather simple cases and how craps can effectively or simply represent them in a very simple way. The first case I would look at is related to player focus. What makes a player tick? What makes a player tick when it really matters. If I can plot a quantitative relationship, high stake games versus performance, say, the way in which a player plays, and playoff games or regular games, if I can effectively represent it and use it for my analytics, that would be a boon. I want this identification process to be very simple. I don't want to look at elaborate slides, lots of data analysis and data dumps. I want a concise cheat sheet kind of representation. More specifically here, a rod who used to play very well in seasonal games, but was bit of a choker when it comes to the playoffs. A lot of very specific training elements were incorporated by coaches to simulate pressure conditions, etcetera, so that he also ticks off in the playoff games. Let's look at how a graph can be used to represent this situation. A very familiar structure, so to speak. Nodes. You have edges and then you have relationships. You have the main actors here represented by the notes, vertices, the coach, a rod himself, and the playoff games, and how the coach, the edges, actually represent the relationships. The analysis, the simulation, and the actual action of playing a graph says a million words. The next case I would look at is based on team focus. Do you remember the Chicago club's strategic victory in the 2016 World Series? How they overcame Cleveland's strong pitching and corrie clubbers dominance. So this was again a relationship Analysis which was done, you know, on how to tire the opposition in this case. And then what relationship did increase pitch counts in the beginning have again, this can be effectively represented by a graph as seen here, like how the pitch count and the sedition drills by the coach did play a relationship effectively in the gameplay. Let's go on to a broader view of how this all spans out in the current industry. The graphs we saw are just like tiny drops in the ocean. Knowledge graphs are now the norm of powerful AI based engines and analytic systems. Powerful knowledge graphs are now becoming the norm in any industry, so to speak. Let's now go down to the underlying technology and underlying terminologies used and the preferences of concepts and notations. Why do we use Cypher Ql? Cypherql is a very effective, concise and text based representation. So it's in fact text based visual art, so to speak. So, you know, if you actually look at it in this particular case, the relationship over here is a person lives in so and so city and knows this person. So look at the way in which it's effectively represented. This graph. It's truly, truly very simple and very clean to understand. So you want your representation to be not confusing, not ambiguous. And I feel cipher QL is the perfect choice. There are other choices like gremlin, which can make it overtly complex. But definitely in terms of the visual art, the clarity in which the nouns, the relationships, you know, the verbs associated with other nouns, so to speak, are represented is very effective in the Cyberku L format. Let's look at a 10,000ft view of how we want the solution landscape to be here in a cloud native ecosystem. At the very heart of it you would actually see a neo four j instance. A neo four j graph instance which is used for storing the graph data. Our favorite go based lambda, the high concurrency and high performant go based lambda which is used to actually drive these analytics modules. Neo four j database instance would again be deployed in an AWS cluster. And then as far as the scalability goes, there are some claims that at least as per the case studies published on the neo four J website, that it scales like 1000 x faster than your SQL for greenfield development and building systems ground up. Having looked at the 1000 10,000ft architecture and again stressing on the usability, the actual use case, this is going to solve the sports peaks analytics users both on mobile and web browsers. There are a host of other tools used here right from monitoring, storage, etcetera, you know, also ensuring that the workflow does authentication, role management, so to speak. Going on to a few code snippets. Hey, one thing in the previous slide I forgot to mention one part, you would see a small lightning bolt. Next to the Gopher bolt is the actual driver used to connect to neo four j and user query in CypherQL in the Golang ecosystem. Coming onto our favorite part, a few code snippets. This is a simple code snippet for the data definition languages or creating the model using Neo four g. A bunch of create and update statements used for representing teams, useful representing players and what kind of relationship the player had. And you know, a few other attributes for the player have been mentioned here. It's very simple. It's just like boilerplate snippets to just give you the idea, more ground level code and go lang awesomeness. In the true spirit of simplicity of Golang, just a few lines of code to actually connect to the neo four J driver. Again, to reiterate, the name of the driver is called bolt. Just some basic authentication at play here and then a little more awesomeness here in terms of a basic insert statement in order to insert data corresponding to a, an item onto the database. This could be a player, it could be any entity per se. Going on more further, a little bit of deep dive. One of the sports my daughter specifically likes, it's lacrosse. Lacrosse is like a fun combination between your badminton and, you know, like butterfly catching and I would say a bit of hockey in some way. So a lot of graph based relationships can be used in lacrosse is what I realized over a period of time. And this is again a snippet which talks about a few intricacies, about, you know, how some basic level of identification can be done using relationships. And graphql. You would look at this particular idiom here. It's a low submarine shot and a rainbow pass, which are pretty tactical moves in the lacrosse space. And also you have a bunch of other paradigms here like turnover and save. That's again neo four j model model used to set up data for this particular regard. There can be some advanced functionalities here which can be used from the graph to actually, you know, analyze in greater detail. One such functionality is the cosine similarity. Rather a mathematical functionality is cosine similarity. So using a graph database enables us to calculate and compare cosine similarity between two nodes. In this way, one can compare opponents, games, players and other nodes against each other to better understand strengths and weaknesses. And how to make the appropriate adjustments. With 30 teams in major leagues, there are bound to be programs that play similarly to each other from a data science perspective in aspects of the games that may include type of relief pitching or stolen base percentages. Evaluation of your opponent based on similarity allows for anticipating given game scenarios, which leads to better practice and strategy. More graph based concepts. You have a bunch of correlations and indices from the graph world which can again be used, you know, combined effectively with go as demonstrated by the few go snippets to actually come up with advanced analytics. One more aspect which actually shows the significance of why sports betting is such a big deal and analytics is the backbone of sports betting. So the daily fantasy sports was red hot in 2014. About 1.5 million Americans paid more than a billion dollars in tournament entry fees and FanDuel grew about 300% in active customers. So you know, there are millions and millions in dollars pumped in by high stake companies like ESPN and Disney in the entire betting world. And you know, Golang, with all its awesomeness, can be used in this ecosystem. A few useful links on this regard have been mentioned here in terms of the repository used and the code snippets which can be used for further analysis. Finally, to wrap this up, a retrospective on what this presentation was meant to be and what this was not meant to be. This presentation is a blueprint, a thought provoker on how sports analytics can be clapped with the awesomeness of Go ecosystem. To achieve advanced analytics, there is an indirect emphasis on using a particular tech stack, like sticking onto cipher QL, which is like a more easily representable format for graphs, and then also teasing the thought of how this can power knowledge, graphs and AI's. And maybe on the other end of the spectrum, this could also be used to power tiny devices like tiny graphs on tiny go. Who knows? That's it for this presentation for now. Hope you liked it. Please do drop me a note on the feedback. Thank you so much.