Python Beginners Project | Predict Cricket Score | LIVE Project Building | GeeksforGeeks

Key Takeaways

Hello. Hello. Just a minute, guys. Just a minute. Just give me a minute. Okay. Let's begin. I my audio is clear. Please tell me in the live chat if everything is fine. And I'll just check on my end also. All right. So, hey everyone and welcome back to Geeks for Geeks. So, guys in this video we are going to start with a very fresh machine learning project. So, it's a machine learning project by the way, not a deep learning project. It's a machine learning project. And what tech stack we are going to use in this project, let's discuss that. So, in your screens you can see that we have a very uh okayish UI, right? It will work, all right. So, this UI this design this website is basically in React JS. And our back end would be running in fast API and our machine learning model would be trained in Python. Okay, so for this we are going to use our cricket data set which we would download from Kaggle. So, this is it. This is a pretty much about the stack. So, if you are someone who is interested in joining a React JS front end with a fast API back end and a machine learning model of yours, you are at the very right place because we are going to do this today, right? So, if you are someone who does not know much about machine learning, I would say please learn that because that's important. I'm not going to discuss concepts like why we use this, why we use this. This is like basic. You might know about all the machine learning concepts. Our main goal is to join everything. I know how to train a model. I know basic HTML, CSS, or React JS. I know uh basic fast API. If you don't know basic fast API and HTML, CSS, that's fine. But, yeah, main goal you you must at least know how to train a machine learning model. We're going to follow the exact steps, but no, I'm not going to explain you each and every basic layman terms. Okay. You can just use ChatGPT or you can learn about machine learning properly to know about these terms because if I go to explanation in-depth, uh this session would be for 4 5 hours, and we have to finish it in 1 or 1.5 hours. Okay. And one more thing, guys, uh after we end this session, the live session, the recording would be available, right? And also, guys, I'm going to provide you the complete source code. It would be uploaded in my GitHub account, so I'll share you the repository link once we end the session. All right? And before we finally start, you can just check out the description. Some links are available here. You can check out all the important links. And uh we have uh lots of links like if you want to contact us, we have Twitter, Instagram, and LinkedIn handles. You can contact us. And uh you can actually explore premium live and online courses by the link provided in the description. Okay? So, enough of talks. Now, let's roll. Okay. So, here you can see that this is a React JS front end. All right? So, our main goal is what? Our main goal is to train a model. To train a model to which we're going to provide these details. Okay, suppose our model is trained now, and we're going to provide some details like batting country, bowling country, and the place or city where the match is being played. Okay? Because the weather map matters, so generally the city would be dependent. Okay. So, if it's like India plays well in India, some in some other country, India might not play well. Okay? So, that's that's the thing. Okay? That's why city is important. Okay? And our data set is also loaded with these things. So, yeah. City is really useful to determine whether you're going to put you're going to score really good at this place or not. Okay, so city is important. Then the current score and how many overs are done and how many wickets have we lost. If suppose we have lost nine wickets. Okay, suppose we have lost nine wickets. Suppose eight wickets. Okay, and it's 25th over. Current score is 250 and like run scored in last five overs is close 35 suppose. Okay, let's predict the score. Uh right now I think I've turned off the model, but accordingly it would predict the score. Let me just refresh and let me just check if my model is running or not. So. Yes. Okay, so suppose this is a match and suppose the current score is 250. Overs done is 25. And let's let's predict on lesser score 120. Overs done are 25. Wickets lost so suppose we have lost eight wickets and runs scored in last Okay, so let's say 32 runs. All right, so if I click on predict, you can see that the final score would be 250 max. Okay, so maximum you can score 25 like right now we have we're in the 25th over. Maximum you can score is around 250. Obviously last people are generally not that good players, but according to the place situation and according to the previous historical data of this specific city and India versus Pakistan and also the remaining remaining overs, the score is predicted. Okay, and obviously the model is not 100% correct. Okay, we have trained the model to an efficiency of 60% around, but yeah definitely a good model is above 90%. So, for that we need more data data more more data set and also more training time. Definitely we can do that with more data sets. But right now, I have picked the data set and trained it with very less time so that I can take a session for you all guys. Okay, so let's continue. So, this is it. This is These are the details which we are going to provide to our model and model is going to predict the score for us. This is the main agenda. Okay, so first guys, we are going to focus on our model. Then guys, we are going to focus on our back end. Okay, then guys, we are going to focus on our front end. Okay, so this session would be divided in three parts. Model, back end and front end, right? Done? Okay, so let's do it. So, let me just close this and let me just uh jump to the main agenda. Okay, so here guys, I'm going to create a new folder, very fresh folder and I'm going to call it as Crick score. Okay, easy. So, Crick score. And uh guys, this Crick score would contain two major things, front end and back end. Back end would contain our model and back end. Okay, so let me just rename it back end. All right, back end and then guys, we are going to have another folder named as front end. All right, so for now, let's focus on this back end folder. So, back end folder would be containing two things. First thing is our model training part. Second thing is our back end where we are going to create APIs which connect the front end and our model together. Okay, so let's do it now. Hi everyone. Hi Saurabh. Good evening, Gautam. Yeah, thanks for joining and definitely you can share this with your friends who are kind of trying to build a college project or something like that because you can assume that yeah this would be a really good major project which you can use in your colleges. Okay, so check this out completely. Okay, just if even if you don't know anything, just watch the tutorial completely so that you at least know okay that's how things work. Okay, it would give you an interest to learn machine learning or web development also, right? So yeah in the back end part we are going to create things step by step. Okay, step by step we're going to create multiple feature extraction files. First of all let's talk about the database. Very important thing is the database. Let's talk about that. So database is very simple. If you just search for cricket score database and you would get lots of data for it. Like cricket is something there are infinite amounts of data for it because every day some match is played, every year lots of matches are played and well lots of popular countries like India, Pakistan, like Australia, different different countries actually participate in it. Okay, so we have lots of data here. Okay, so we have cricket data then we have cricket data set like I actually picked this one I guess. So let me just check. Yeah, we actually picked this data set. Right? We picked this data set and like it's by this author Thirisan. So shout out to this guy. And you can actually get the whole data set. Here we have complete details. Okay, complete details you can check out. Now I'm going to go to nine files and here we have CSV file also. You can check out this one or if I check out Just a minute. Let me just confirm. Okay, then we're going to continue. Okay, I found the exact data set which I actually used. So guys, this is the exact data set which I actually used. So you can just search about this retro sheet for cricket. Kind of it's like Excel sheet filled with who is batting, who is bowling, current score, things like that. Okay, so yeah, you got to go here and here in the data explorer, we have different different types of formats. We have CPL, we have IPL, we have 20s, we have T20s, we have ODIs. So since right now we have a match between India. Okay. So like we are having ICC ODIs, right? So we are going to maybe download this one, right? So if I just go here, you can see it has all my AML files. So we are going to convert this to a readable format. We're going to do that. So let me just download this one. Okay, so if I download this one, it's going to be downloaded. And let me just click here. I'm not sure why it's not completely downloading. Let me just click again. Um Let's try again. So, if I click on ODIs, okay, and if I click on download, it's not downloading just ODIs. So, let's download the whole, okay? So, it's just 166 MB, so there isn't any issue, okay? So, let me just uh bring that downloaded archive. So, guys, the I have already downloaded the data set. So, here you can see. This is the data set which I downloaded. I pasted it in my back-end folder, okay? So, I'm going to just extract it. So, I'll just extract it. And again, it's going to take some time. So, instead of waiting this much, I have final extracted file. So, let me just bring that. All right. Just a second. So, I'll just delete everything from here, and I have the final extracted file, which you already downloaded, okay? So, again, it's going to take a few seconds. So, let's wait for it. Okay? So, till then, what we're going to do is we are going to open this with VS Code, okay? Our back-end folder with VS Code. Let's do it. And yeah, you can see it's coming here, and it has all the YML files. So, it's going to take some minutes, okay? Maybe 1 or 2 minutes. Let's wait for it. Uh till then, what we're going to do is we're going to create a Python file, okay? We're going to create a Python file here. And let's name that file as feature extraction. Feature extraction. Okay, feature extraction.py. Okay, feature extraction.py. Okay, so this is the file. Um actually, uh let me just do one thing. I'm just going to cut it and I'll just add it to my model folder. Okay, so that there is no confusion. Okay, so yeah, this is the feature extraction file. And uh here, let's wait for this to be completed. Then we're going to call the file. Okay, we're going to call the file. Till then, guys, what we can do is we can make a few imports which I'm going to directly use so that time is not wasted. So, we're going to use NumPy, Pandas, YAML extractor, and then OS for file management, and then tqdm for getting the percentages, and pickle for saving file frequently. Okay, so that it's a huge data set. So, we have to save things frequently so that we don't need to start things from scratch every time. That's why we need pickle. Okay, so uh make sure you have installed everything. If you haven't installed any everything, just type pip pip install or just create a requirements file. You can do that on your own, right? So, yeah, these things are required. And uh you can use a directly uh use a Jupyter notebook also. That's also fine. You can do the same thing in Google Colab. That's also fine. But in my case, I feel like uh VS Code works for me, so I'm using VS Code, but it's no compulsion you have to use VS Code. You can use any uh notebook also like Jupyter notebook or Google Colab. That's totally up to you. Okay, so yeah, I use VS Code a lot, so that's why I'm going to stick to it. Okay, so I think that now the copy part is complete, and here you can see all the formats are here. Out of which the main main one is ODI. So, I'm just going to cut it, control X, and uh I'm going to paste it inside models. Okay, and I don't need these archives now, so I'll just uh delete it. Okay, it's going to take a few seconds, and you can see that's deleted. Okay, so this is my data set, So I'm going to call it data. Okay, I'll just call it data. So this is my data set. It contains ODI results, okay? Like everything. At which ball who is scoring, at which ball who got out, something like that. Okay, so this is it. Now we have feature extraction file, okay? So here in this file we are going to simply call what? We're going to call our data directory, okay? And we're going to just extract all these and extract all file names here. Add all file names here. Okay, so like if you want to print it, let me just print it for you. So I'm just printing first five file names. Okay, so it's going to say that system cannot find path data. Okay, so it's not able to find the path data baby because I am here. So I need to go inside models folder, okay? So open an integrated terminal and now guys I'm going to run this and you can see that yeah, you got this. Okay, so it's like the backend folder does not know where data is, okay? So yeah, it's running according to the models folder, okay? Make sure you're doing that. Make sure about the paths, okay? Don't get confused. If I'm running it for the backend folder then I need to just say dot slash models slash data, okay? I have to write it like that. But I am inside model and I'm inside feature file. So I have to look for data, so I'll just look in the same folder, that is models. Okay, so now we have the file names. You know everything is clear, data is loaded, easy. Now we we are going to just say that uh like we're going to read everything, okay? We're going to read everything. Like it's like I'm going to read everything and I'm going to skipping I'm going to skip the empty data, okay? I'm going to read everything and I'm going to skip empty data, okay? So it's a for loop which is going to read everything, convert the file YML format to a readable format. You can see it's converting it to a JSON format, okay? So that we can read it, okay? Then after that we are going to just save it, okay? We are going to just save it, okay? So we basically got all the data frames. You can see we have got all the data frames, all the readable formats, right? And we are just saving it to final DF variable. And I'm just making a copy of final DF out of here. Okay, that's it. So, it's going to take a few minutes maybe to actually run this part. So, that's why I'm not running right now. But yeah, it's going to take a few minutes. Okay. So, now guys, I am going to drop extra columns. Okay, so in my code, there are some extra columns like metadata version and so on. Let me just run this for you if you are feeling like I'm not running it. Okay, let me just Okay, I did one mistake. I don't need to run it like this. Yes, I can run it here. All right. So, you can see that I'm inside models folder and you can see it's going to take some time. You can see that around 10 minutes are needed, 10 12 minutes are needed. So, if instead of wasting the time, I have the final things. But yeah, this would work. Okay, this would finally contain kind of an Excel sheet, okay, which contains data, okay, which contains data. Okay, something like that. Okay. Now, we are just removing these columns from my Excel sheet, okay. I'm removing these columns from my Excel sheet because I don't need them. So, why to just increase the load? Okay. Then from that, I'm just going to say that in my Excel sheet, I have a genders column. Okay, so I'm just getting a genders column, okay. And you actually don't need this. Okay. I'm just saying that keep only males, okay. I'm talking about male players, okay, male tournaments, okay, men's World Cup for now, okay. Don't be angry at me, okay, but right now I'm just taking men's because I need to just minimize my data so that the code runs faster and model is training faster and we are more specific to men's for now, okay. So, yeah, I'm just saying that from from my from my Excel sheet, remove the men's remove remove the women's women's part, okay. So, it's going to remove all the female players, all the scores, all the matches for them, okay. So, it's like I have scores and runs and teams only for the men's for now. Okay, so that's why. Okay, and I'm just saying now I have this so I'm just going to say that just drop all the gender part. Okay, just just drop after that just remove the gender column because we don't need the gender column. Right now I know I will have only one gender that is male. So I'm just removing the gender column now because I don't care about anything in the gender part because now only know that yeah, in my whole Excel file we only have males right now. Okay. After this code you only have males right now. So we have removed males. Easy, right? Now after this I am going to say that uh like my goal is to target ODIs. Okay, 50 overs. Okay, so I'm just going to say that just get the overs who are 50. Okay, if the match is lesser than 50 overs, okay, suppose if by chance my data set contains a match with 20 overs. It's not an ODI, right? It's not an ODI for me. For me 50 overs is an ODI. So I'm just saying that from my data just filter out only 50 ones. Okay, so I'm just keeping all the matches which have 50 overs only. Okay, not less than that, not more than that. Okay, so this is my feature extraction. Okay, this is my feature extraction and you can call it filtering of my data. Okay, so I'm filtering my data like this. So shortening it and only getting the relevant amount of data. So after this again I'm saying drop these two columns also. I don't need overs because I know overs are 50. Okay, and I don't need match type because I know that 50 overs means ODI. Okay, so I'm just dropping these columns. Okay, and after that I'm going to dump it to a file named as dataset level one {dot} pkl because this much of progress is going to require half an hour or maybe 15 to 20 minutes at least. So you won't repeat it again and again if some error is caused. That's why after doing this much, after successfully doing this part, I'm just saving it to my data set level one PKL file. Okay, so you just need to run this. Okay, you just need to run this like this. And after 15 to 20 minutes, a file named as data set level one {dot} PKL would be created. Okay, so I'm going to directly paste that file because we don't need to waste that much time. So this file would be created. Okay, this is the progress after doing everything. The data the data which we were using is finally saved here with all the filtering and feature extraction. Okay, easy. Now after this, let's work on another file which is feature extraction two {dot} pi. Okay, so now here we are going to start working on data set level one {dot} PKL. Previously we were working on data. We converted it to data set level one {dot} PKL by just removing unnecessary things, just filtering out things. And now we have data set level one PKL. Now here we are going to do again do some operations, get some extra columns, remove some extra columns, and then again create a PKL file so that our progress is saved every time. Right? So that saves a lot of time, that's why I'm creating a PKL file. Okay, so automatically it would be created once you run this file, once you run the first file. Okay, once you run the first file, step one is complete, progress is saved inside data set underscore level one {dot} PKL. Okay, easy, right? All right. Have a cup of tea. You guys, if you have any doubts till now, please tell me in the live chat. I'm just having a look in the live chat. And also if you are worried about the source code, where will be the source code? I'm going to share it in my GitHub repository and also provide you the link in my GitHub repository. Okay, so don't worry about that. Okay? All right. Let's continue. So, let's begin with step two. So, step two is again filtering things out. All right. So, this part is uh This part is a little bit uh easier, you would say, because it does not have that much things. Okay. So, again, I needed some imports. I did some imports. And uh then, I have loaded my PKL file, okay, dataset level one {dot} PKL file. So, whatever progress I have done here and saved it in the dataset level one {dot} PKL file. Now, I'm calling it inside my feature extraction two. Okay. So, now I actually don't need feature extraction one file, okay. You can delete it or do whatever you want. And I also don't need my data now, cuz we have already converted it into my PKL save format, better format. Okay. Now, we have this. So, what I'm going to do is again, I am going to convert my data. Like uh right now, my main goal is what? My main goal is what? My main goal is to get innings, okay. So, it's like uh I'm going to get first innings, okay. I'm going to get first innings and the deliveries out of it, okay. How many balls are left, okay. So, in 0.1 ball, this score, 0.2 ball, this score, 0.3 ball, this score. So, from my small dataset, I need to expand it, okay. So, for that, I have already written a piece of code, okay. You can just If you are not able to understand my piece of code, don't worry. You can use ChatGPT for it. It would just explain you, okay. This guy has converted these columns to these, okay. So, yeah. All right. So, what I have actually done here is, you can see what I have actually done here is, I'm just saying that uh get these things, okay? And create I'm basically creating a new Excel file, okay? Or new data set for from from my existing amount of matches, okay? So, I'm just saying I'm just saying get innings, okay? Get the first innings, get the number of deliveries. From the number of deliveries, I am just getting all these things, okay? I'm just getting all these things, batting team, teams, and ball of the match, batsman, bowler, and so on, okay? I'm getting all of these things. Okay? So, I'm just saying that uh like if wicket, then player is out, something like that, okay? And finally, I'm passing all this data. You can see match ID, teams, batting team, ball, batsman, bowler, runs, and so on to my new Excel sheet, okay? And I'm just saving it inside delivery.delivery_df file, okay? So, if you are going to print this, so let me just print it for you, and I think we can waste 2 minutes here. It's not bad, okay? Uh okay, I think I need to run it. Yeah. Okay, so it's going to take a few uh like 2 minutes at least, so let's wait for 2 minutes, then we are going to continue, okay? I'm just showing you what it actually now contains, okay? What my data set looks like. Hi Robinhood. Hi Santosh Kumar. Hi Cloud. Hi Matthew Zerudo. If you are having any doubts, please tell me in the live chat, and if you want to get the source code, I'll be uploading it once we end this session. I'll be uploading it. I'll be sharing it on GitHub, okay? Let me just do that till now, okay? Let me just do it for you all. So that you want to check or you want write the source code along with me, you can just do that. Okay, just wait. I'm sharing you the GitHub repository. Um All right, guys. So you can see that this is kind of my data set. Okay, that's what actually looks. Okay, you can have a look. And I'm just going to paste this here so that you remember how my data set looks like. Okay, so my data set now looks like this. Okay, it has match ID, it has teams and bowling and batting. The dot dot dot is actually the hidden part. Then the city, then the venue and so on. Okay, somewhat this is kind of the data set right now we have. Okay, after all the conversions. Okay. Uh okay, just a second. I think I need to do some changes. Uh okay, anyways, guys, at the end of this video, I'll be sharing you the source code. Right now, I have lots of load in my PC. Anyways, so you can see that right now, after all this, okay, after the first part, okay, converting it into readable format, then again, after that, converting the balls getting out of getting all the things, okay?