 If you feel like, oh my god, don't worry, it will get proved in very short time. And don't expect to, just to set the expectations right. This is just a fun session, sorry. Don't expect to learn too much, remember the next best data scientist, how do you do this session? So, I've said it in the right picture. This is Galileo Galileo, also classed as the father of water science, named as father of water science by all the registrations, or even best people of the time, many others. For this, can you see what it is clearly on the back? Which is Mancha and Mofiza, once it's okay, yeah, Galileo's father, he has told it. Every time a beautiful brick, a piece came in, the sandal would swing just a little bit. And this intrigued Galileo. He wanted to know the oscillation of that sandal. So he decided to use his own army to measure how the sandal would swing. In order to know how strong the piece was, every time the line, every time the period of oscillation was the same. So if the sandal would get displaced in this much, or if it would get displaced in this much, the period of oscillation was being same. This intrigued him. And then he started his studies with simple formulas. And he did a lot of this, I've seen for a long time, or maybe just before that, he got this idea of using this property, he called it isoprolism, the property of taking the snake on one side of the oscillation. Irrespective of certain factors. So in 1640, he was struggling to use this property to create it. And then he gave out the basic design. But at that point, Galileo had already third line, and he was a house artist for having gone to work in the design. But still it's a period of pendulum that remains the same, irrespective of the initial algorithm, as well as it remains safe, even if you change the mass that you suspect. Here, in 1656, the parts, the laws of parts were just became popular. In two of the 18th and 19th century, these laws were arranged, and they empowered the Industrial Revolution in some sense, because they were used for time keeping, for band-raining schedules, in factories at radio stations, and just making a thing like this. But this thing I just was thinking about how to go about solving this, what assumptions can I make, what simplifications can I make. So whatever it is, you want to find the time which fits this data best. Even in unknown point where you don't have the data, you should be able to predict what the outcome will be. If it will be as close to the real process as possible. That's what he said. He's looking at now what I'm doing. And he's smiling and looking at me. He just puts me one question. What is the length of the video for one second? I want to make it long, right? I need to get along with the video with one second. I say, let's simple this. You just have to draw a horizontal line. You think the surface is equal to 20. Come here, take a take on the mid-classifications. And we should be there, because it's a red line. That's your point. You say, see, see. And he says, can you just want to draw the higher-axis ones for me? I said, yeah, I'm fine. It's behind zero. I'm ready, ready, ready. This moment, this is your audience. He is... Just a little bit more. Just a little bit more. Like every software developer will be saying, when it's just been done, you can expect it to pass. And you can be one more chance. I'll help you with something else. So I'll have to make it what it is. Should I draw it or not? Then no. Can you do a lot of this? Can you do it smart? Should I just make a custom? I say, what? Can we do something else? Can you do something else? Anyhow, I'll do the square and you do something else. So let's do that. This line here, I have a square. And I'll do the more exercise again. The direction is even more fine. Which means I need the... What am I asking? Sorry? Yeah, there's a line. It takes away the energy. I mean, as it is, it is probably worse than compared to what it was earlier. This could be one conclusion. Can I do something else? Can I do something else in this? Can I do something else in the end system? Actually, I'm going to take the square root of 10. And let's see what happens. Oh! You've got an intersection on the positive side. Okay. Really. 0.5, 7.0, 0.5. It didn't be accurate at all, but it turns 0.5. So this is the square root of the length. You get the actual length, you just need to square it up. That will be close to 0.25. And 25 centimetres, that's the length of what's going to be when you'll get to one second immediate. You go and check the class. There probably haven't been any questions in the long. Let's see if I can... This is the world's best friend. He took me to ask me. Is there a lot of questions? The way we do this is to disturb the ball in the way. So we'll let it settle. I'll start my talk. Watch once I see that you've said one extreme. You've got 20 oscillations. It goes once, comes back, that's one. Okay. You've got 20 oscillations. Then you stable. Start 9. 15. 18. 19. What are the ideas? Very good. Okay. This is 51.8. Good. Now we'll have to use our intercepting. So, how much is it? 51.8. 51.8. 51.8. 21.8. 21.8. minus 11.8. So, 1.29. There it is. Square it up. Watch the view. Square it up. Okay. 1.664. That's one thing we need to be honest about. There's going to be some measurement inactivity. Because the kind of devices that we're going to use to measure the level, the level of the light, is just as simple as a day. And we say that the error is likely to be close to, let's say, 1.5 centimetres. It's likely to be less than 1 centimetres. Now I can say it's 5 centimetres. That way, the time again, the size is going to be slightly more than that. Including the pitch, it's better because we've got a mark for such that the pin moves also. And that's where it comes on the table. I probably am not behind the surface. Okay. Let's measure better just a bit first. This is 1.62. 1.62. Okay. Plus the radius of the ball. Because this is the centre of mass. We'll likely be able to assume that it is at the centre of the sphere. So this is a standard that is more or less. So this is 6.54. Is that the size of the ball that is more? Let's see. Whatever it is. We'll just take an estimate that it is around 5-6 centimetres. How much was it? 1.62. Plus 3 centimetres. So that is roughly 165 centimetres. This is 1.66 centimetres. So this is within the exact accuracy. Any questions for Misha? Yes. And you talked about the way you expect the ball to be very, very heavy. You get a larger length for what it is. Yes. How is it? So if we can do a calculation of that, we will be able to give you a very good question. What's the nature of how it is to open the net inside the net? A liquid, yes. Sometimes what happens is that most often we get carried away by the power of the techniques that we have. And for these days, we have got a life of computational techniques, numerical techniques per se, rather than going beyond that technique. So if you look at that differential equation, it is possible to solve that differential equation by making an assumption that theta is small. For small theta, the last number of times theta by theta is 14 equal to 1. So if you substitute that, you get the equation of simple harmonic motion, the one you get from spring object just when you spin it. So that equation is easy to solve, and then you will come up with a formula. So that's the formula here that you can do by solving it. So here it is very, very obvious that this proportion 2 pi by 2g is your advancement. So if you see if you just divide this 40.8 pi by 0, then that's the problem for any advancement. It should turn out to be 2 pi by 0. Again, the intercept value, it's a very small value. You should forget this value if you come up with a thing. It's because of my observation. It shouldn't be really zero in the real mathematical relationship that exists. So these are the things to fit with underrated these days. Earlier on, that was the only problem with the case. That's what I want to take. That's what I want to take. That's what I want to take. That's what we come back to. So let's try to write the equation for me here and then, okay, or actually just write it here. What we have is this thing, simple equation k0 plus k1x, right? Simple line straight line. Let's try this problem with me. Transform x to the square root of x. Okay? So now, this is actually a solution. What we find is that k0 is really very small in the set curve. It's really very small. So what we do? In ideal world, if you move this over for the vertical set of observations, it may be k0 to the larger of 0. Okay? And you'll get a relationship like k1 times square root of x. But try to remember that this is again immediate for 20 observations. So essentially, you want this is going to be equal to some constant. This, 2 pi instead of k, this is essentially a proportionality constant. So you found the relationship between that and me. Yes, it's a problem. Yes, it's a problem. Find that. You want something about whether it's a square root of c, or a square root of r, or another root of b, but I'm fine with it. So that's why I'm here. I'm not even writing and I'm fine with it. I said, I'm fine with it either way. Okay. Well, we've raised the question. You're interested in how do I come up with this square root of x? What is the other square root of whatever it is? Yeah, how do I come up with it? So this is where I stand up because a little bit of light. So you have to have some experience to be able to, you should know one of the candidates, who can get the answer. With your confidence. And I would say, in fact, you can do similar things. You can do a bunch of different polynomials. And see which gives you the least of this square root. What are the different polynomials? Do you have any other questions? Yeah, we can have a problem, especially if we have a problem. We can have a problem. So that's why I'm here. Good. Any other questions there? Yeah. So, when you look at formula, for the kind of how it works, how it works, how it works. So, if you use that formula, you can see, there's a difference in the equation. And again, you see the equation. You add an error to put it in the formula. That is the power of data science. You don't have to worry about the actual formula. But if you utilize, for example, in this case, because you're adding an error to it, if you will more or less generalize and provide an answer to despite rules, if you have a complex formula, you don't have to worry about moving the polynomials. Okay. It's also high-level. It's in 5, I started with this. It's also high-level, typical variance. Like, if you are not looking at other statistics, other important figures as well, how good the critis, does the data actually not flow away and get this rate line. Okay. So, these are the factors which are much different over the which is called a high-level. So, you can't just depend on the analysis that you have at the set. You have to add some sort of some sort of judgment when you're using it. Okay. I don't know if you add because it's too good for you. It's all that different situation. Invariably, somewhere, in your equation, why do you fight with this grid with height? So, with what is it related? Why do you fight with all the possibilities? I can't take that. It uses partial data, to find out what's the minimal function of the error path. Right? So, you write the last function which is associated. So, this is called the answer. However, do you remind me? Well, sort of. I got it? Yeah. I'll give you 5,000 million.