Everything about Sports Analytics | Data Analytics in Cricket | IPL Fantasy League Prediction

[Music] [Applause] [Music] uh hello everyone we'll start the session exactly in two minutes uh let's wait for more people 14 people watching let's wait for two more minutes and we'll start shwaytang is also with me and i'll also call his name very quickly just give us two minutes in the meantime in case you have any requests any any any suggestions or anything you want to have just reply on the live chat itself so that i can see in case you need some specific uh sessions on any for any topics probably on pca on feature engineering or any any kind of topics just let me know tableau power bi any specific eda classes or anything be it on anything just just try to write down in the comments itself and we will get back to you very quickly just uh just a notice to you guys next weekend's class is going to be on google analytics so i will have prashanth who is one of my colleagues he will be joining for that session so yeah in case you are new to this channel please like share and subscribe the video and we'll exactly start in one more minute we'll just wait for one more minute so that people join in uh shaytan can you go on full screen mode please thank you i should thank you there you can you go on full screen mode for your presentation yeah thank you sritank just give it one more minute swear thank we will start quickly yeah no problem okay so who all are there can we have covet 19 prediction definitely manish we can have it i will keep that in radar probably next to next weekend's session is free so we can have something on covet 19 coin ending kind of prediction or kind of a dashboard where we can we can scrape data real time from twitter or something maybe we can see what we can do uh hello everyone hi hi hi shankmesh thank you interesting yes could be interesting i'm really excited for sports analytics because the moment i saw some post on linkedin related to sports analytics i quickly reached out to swathank and i asked for his time and luckily i got his time so we will have swaytank he will be explaining everything about sports analytics and he will also explain you some of the use cases related to cricket and of course there are abundant use cases will not be able to cover everything but yeah if you are interested further probably let's say you're interested for football or some other sports if you want to do that kind of analysis just let me know and i will have shaytan with us for some of our future sessions and we can conduct a conductor session quickly anybody has any doubts 31 people watching please like the video guys i have been following for thanks post on plankton oh nice sangmesh thank you thank you please like please guys like as much as possible because uh based on the likes itself facebook sorry uh youtube will recommend the video to other people as well so the algorithms have changed so it's very difficult to get traction very interesting thank you roshan okay in the meanwhile let people join in i'll introduce myself people know me but hello everyone this is satyajit patnaik and welcome to my channel in case you are new to the channel please subscribe the channel and hit the bell icon to get notified on my future videos in session today we have with us dr patak so who works as a data scientist in the sports domain is also a physiotherapist at center of sports science and he's also a part of our various data science communities where he helps other fellow members and data science enthusiasts as well as most of you already know i have recently started doing live workshops and webinars every day like every week sunday at 7 00 pm ist probably 7 to 8 pm it depends but mostly most likely it will be 7 pm isd so today's topic is going to be on sports analytics and we'll see how data analysis can be done in sports domain at least i haven't tried it so my this is going to be very fascinating for me and uh svetang we also talk about some of the use cases in the cricket areas like what can be done in the cricket be it a fantasy league prediction or be it a cricketer prediction cricket analysis players analysis all these kind of things that can be done right uh so let's get quickly started on this in case you have any questions in the middle of the session you can post it as a comment or in the live chat and i and we'll try to get back to you with the answers if your questions are somehow missed you can definitely reach out to me or svetan directly links to our linkedin profiles are already shared in the description of this video you also have my whatsapp numbers as well so you are free to ping me anytime just a disclaimer like the last session like like the last few sessions uh shaitan has created presentations and quotes which will be showing live to you directly during the session but this is only accessible to you once i get 100 likes on this video now that's not going to be demanding but just a request to like share and subscribe my channel and trust me guys will come with a lot of free webinars and workshops on various topics related to this niche okay so let's get started uh should thank over to you guys over to you thanks sorry uh you can probably get started with it okay sorry i'm sorry i had me ready you can probably start again hi everyone so thanks for joining this video okay and i would like to thanks him also for letting me know about opportunity to have a give up video about the sports over here so when we talk about sports so there are various sorts of game okay so it's like a big data when we are talking about these things okay it is not confined to only doing any fantasy league or any football and analytics and all those things so whenever i'm working in the sports domain i have to look after the fitness also and then some performance also and when we talk about the performance it can be prediction of how well you're going to play all those things comes in the picture we talk about sports it is not new topic there was a movie on this also moneyball fine and there's also book you guys can go and read that book it is interesting book when you go forward with this topic okay so the agenda of this session is to know about what is sports analytics how it works what other things are over there and what are things we can do what is the benefit who are the stakeholders over there so you can see over here okay so basically sports analytics is umbrella term okay so you you're not only confined to doing only only friction of scores enough runs or how many goals the they will going to score it is bigger than that okay so when we talk about uh uh any sort of fantasyland so you can there is a betting going on over there okay so there's lots of money involved over there similarly uh and during auction okay so we have to understand the market we have to understand how so we have to know what is the level of performance how they are performing what is the fitness level all those things comes in the picture when we talk about the transfer of a from one into another hour's past performance and everything so it's not only about talking what will be the core of the match who will win is far more than that in today's we will be going through a python go to as well and so how run prediction is going on the sim model it's not so sophisticated over here and similarly with the help of excel i will be giving a overview how to measure the impact okay supposed to players are playing so both of them scored a similar thing so which is performing well so in that respect we have to give a weightage over there okay all those things i will go in this session so when we talk about understanding the sports market okay so there's involvement of stakeholders okay franchisees are there they will be investing lots of money by buying any or you can say marketing all those things involved if you understand well enough the sports what is the emotion over there okay so all those things are important suppose if you understand if you want to understand emotion over here so we can go for the nlp some commentators are doing any sort of commentary you can go for the extracting those data from the website or and you can do analysis on that aspect therefore you can know what is the sentiment over there okay so whether they are saying in the favor of a team or against the team favor of the or against a player or nullified all the things can be done similarly each player was performing well in the past oscillations and everything all those things in the picture when we talk about sports market over here because suppose if you want to buy any player okay and if you're buying for x amount over here and if that player is not performing well for your team and that x amount is going to be of no use so all those things comes when we talk about understanding the sports market and audience okay engagement of audience how you are doing okay so all those marketing stuff everything will comes over here that's why with the help of a basic stats or you can say applying some statistical test over here you can analyze to a group of audience or to team who is performing well and all those things comes over here and through that we can know which which here we have to invest on or which team we have to invest on which ground all the things comes in the picture suppose if you are playing a match in some area where the ever okay traveling is difficult then again the audience reach will be difficult all those things is in the picture when we talk about how the marketing sports market work so that is one of the critical area we talk about the monetary expense over here okay similarly this comes the fantasy sports and that is well versed over here okay everybody is uh i think any sorts of fantasy sports and if you know how to predict anything in the future with some sort of good accuracy it will be better is it valid or reliable that is a question mark over here suppose this team is playing a 11 player today and that team is not playing same player in the season the whole the mathematics goes wrong over there because you took a feature of that player while predicting anything it's a dynamic thing over here it will keep on changing all those things all those things are dynamic so we have to take into consideration of the agent which players are playing all those things you have to take care of then we can say yeah that will be going to perform well or not that is one of the thing over here yeah fitness as being a working as a sports scientist and i have to look after the fitness also and fitness is highly correlated with the performance okay suppose if you are if you want to the sex you need to have a good power in your upper body otherwise it is quite difficult to hit long sixes it's not about bulkiness if you think that bulky body will hit far it's not like that um much more is involved physiologically so all the things we have to take care of fitness we talk about fitness there are various system in the body all the system you have to take care of it's not only about being one specific question we can go up it is fit or not fit so we have to all the aspects of the testing over there that is quite needed similarly understanding the sports okay if you don't know xyz of the sports how you are going to work on that sports if you are knowing about cricket only and if you want to go for the basketball or your football athletics any sort of and whether in the long jump if you want to say the person will be jumping far more if you don't know anything about that it's quite difficult to analyze all those things so over here i have divided into the four corrections okay everything is related to each other we can do a team analysis over here as well here analysis as well what is your team fitness okay when we talk about we can group together and we can say fitness what's good or bad all the things we can say similarly individual players if the player is unfit that is going to impact the performance and again your collection of that suppose you're selecting some player and who is not so well who is not so fit again there is high chances that he will go off in the match so these are all things we have to take care of when we are doing all those things although it is quite difficult to get a data the fitness data is quite difficult and it is confidential so we can't get it but somehow i'm working in that domain so i know uh then we talk about match okay so what was the score yesterday for every performed all those things similarly we can go for the season what was the season score last year 2020 or 2019 all those things the historical data as all we know you can say something about the performance over here how the team is performing how years are performing basically in the point of view of match only okay we are not talking right now about fitness okay it is quite difficult to go for the historical data to say about the fitness in the present okay our body changes regularly we can't say about that so when we talk about fitness it has to be present scenario we can't go for the historical data it is quite difficult if someone is running poor or eating something bad it's obviously it is going to cause an impact in the fitness all the things comes over here in the fitness we can't go for the historical data in the fitness point of view it's quite difficult to predict the fitness in the future because it's not so reliable over here uh when we talk about the performance basically your team performance or you can say individual performance how many runs the team score or code in a season or in a match all those things we can go we can do that with the help of simple linear regression okay and yeah as i told you understanding the sports so how to do that all those things so when we talk about their two factors i have written over here intrinsic factor an exchange effect intrinsic means the things that is inside your body okay if these things are not so good again it's going to cause the impact of your body and again all your selection of the suppose you're choosing any player while doing selection that dream 11 and somehow you due to the dehydration or anything physical or physical fitness it came out of the match the whole that point will go zero over there or it will be two or four and it is going to cause an impact on your overall score over there in the dream 11 so yeah intrinsic factor let's play a role and we have to take into consideration that only when we talk about signing factors i will talk to this in the next slide for the changing sectors and how it works although we can say climate which is when we are talking about the cricket only climate we can say which we we can say or in condition we can say or you can say this all the things mainly and opponents okay it's opponent that also comes in the exchange exactly when we talk about that okay if someone if suppose for example without police batting against some new ballers the confidence level of that border will be not so high when we talk about psychological aspect over here so those things will come he will be nervous and all those things will come into the picture over here that's why all these things in the picture again it depends upon how you are training all those things and how your psychological mindset is so what are things we can do with the help of data science over here we can compare two teams yeah we can do that we can do simple descriptive stats over here know about the individual theme stats or overall individual player or team all these things can be done we can undo some sort of correlation uh whether some those things and we we have to do a pc also over here simple component analysis because maybe too much of attributes and you have to reduce your dimensionality over there so all the things comes in the picture and it is there are research people also regarding the pca you can go and search and you can see how pcs are working in the sports domain similarly when we talk about machine learning we can go for the prediction of injuries okay we can go for the injury prediction one thing is that sorry then we can go for the prediction get prediction we know loss prediction all those things can be done the help of deep learning and video analytics uh we can go for the analysis of suppose heat map a boiler is balling a same ball in the same place so that will going to generate some sort of balling uh heat map over there so that in the next session we can analyze those areas and we can train our batsmen how they can go forward with that so all those things can be done similarly in case of batsman we can know nick is there his favorite sort is cover drive or on drive whatever it is okay and we can analyze all those things and we can work those aspects how to move forward okay and if we know some boilers are pulling only over and he can drive the ball so yeah so we can select all those things again when we talk about over here video analytics comes in the picture there are various tools available to do that also so so when we're talking about analytics over here fantasy sports okay so what are things we can do and how we can do there are various ways to do okay one is your weightage method one is your regression one is a monte carlo simulation okay suppose in a dream 11 you when you are doing you can see captains are giving some weightage and vice captains also given some weightage okay so that uh if someone is choosing that captain over there and if they score something or cricket system will give an extra point over there similarly in that basis we can do our analysis also okay you can give some weightage to the runs somewhere it is to the sixes six how many six you can give extra witness to those point because those are the some quality if some batsmen are scoring seven sixes uh that is not a joke you can give a wittage over there we can work with the weighted system uh coming to the today's topic i have done one regression analysis i will show you that over here similarly when we are using any top any here over here we can go for first of all we have to analyze the environment where the match is whether the the which was high scoring or low scoring if the pitch was high scoring we can go for the extra batsman over there if the pitch is low scoring we can go for the baller or we can make factory either way that is how it works when we when we going to select that in the match similarly it's quite difficult as i said to predict the future because there's lots of variation in the body as well in the dynamic it's too dynamic how much anyone can say that they can predict anything but it's too dynamic over here okay so how to do that these are these are the features i have written when we are going to consider the win loss of a team we can consider the result of previous matches okay we can uh match type d match night match probably we are talking about a t20 it has to be a night match only followers in the team we can go for the ssc ranking over there how they are scoring how they are performing over there first how many all rounders are over there uh ssc ranking as i told which i have already mentioned climatic and condition standard okay so when we talk about we can build a model like any classification model and we can go for uh doing whether the team will be going to win or lose depending upon this data again we can add many attributes and just for reference i have given mention over here it's all the attributes are fine so suppose if you want to say tomorrow is a match and how much that they will going to score now we can do that with these attributes we can take average strike rate how many fifties hundreds in the historical data they have scored and with that we can build a model and we can say in the future how many scores they are going to score similarly for the wicked we can take average economic rate how many five wicket hall a strike rate what is a balling technique what is the heat map of the boiler similarly what is the heat map of the batsmen how many scores in the wagon wheel how many score they are scoring in that second wheel area all those things we can take into consideration over here when we are predicting the vein how many runs we get and fitness also we have to look after many things when we talk about fitness i will give a brief idea about that okay so when whenever in in cricket perspective only so arabic i guess many of you have heard about yo-yo taste okay so your test is simply a cardiovascular endurance test okay so they have to under 20 meter distance is there so they have to run to any flow and there are various levels in that starting from five five okay higher the level that means the player is much more fit and depending upon the cardiovascular enterprise over here so that is better so the person will not going to get tired easily that is the importance of that your taste over here okay and if you if you think that the person is fat and they will be having much more fat in in their body and they will going to get easily so all those things we can stay and we can relate to with the fitness and later on with the performance and when we talk about all those things we have to take into content what is the pattern of the team whether they are winning or not because pattern is placed like a psychological impact okay you can say about australian team in the 2003 world cup they won so many matches attitude and pattern only they might mind it what was like that only okay so all those things comes over here in the picture so over here i will be discussing about things and when we go for the walkthrough okay so one is measuring the runs okay let's show you and how to measure the impact of the first of yesterday match only when we talk about blood scored and radio versus code who was the much more having a much more impact definitely pull out over there wickets also over there in the match okay all the things comes in the picture and that and he is making the win also okay so all the things are there now how to do that okay so now i will be i think it's slideshow is fine now so i will be going for excel sheet right now then i will go for the within code okay uh hey shwetak i have a question from rocky he's asking that let's say you're creating some model how uh how do you use the prediction of the toss in the model is does it plays a vital role as a feature or how is it suppose if you are building a building a model and we are taking a consideration of us okay so we have seen in the history that most of the team winning the match with the toss they are electing the baton ball okay so depending upon that only and we can build a new model and we can see for the weightage over there how much weight is mathematically when we talk about but going for the historical point of view and all those things it's like a guest guess estimate only but when we talk about mathematical model okay so we have to take into consideration about the weightage over here how much of it is the toss is giving to you in your model but i would believe that first please okay depending upon the pitch suppose there is a match of high score okay the which is high scoring yeah you are knowing okay and you won the question you went for a batting okay and you scored 200 appearance suppose again the batting second team has a psychological impact over there which is that's how this is how you can say about that because it's playing a vital role over here yeah yeah yeah so next i will be sharing excel suit over here okay give me a moment yeah i mean sports analytics is very it's a very vast domain guys so uh this is for the first time probably in youtube i i haven't seen any stream related to sports analytics so uh this is the very first session so let's see what uh swathang has to show you and yeah from the next class we can probably take a very specific project and then we'll talk on that so that will actually clear out a lot of doubts it's basically a generic topic that we're discussing today yes in this domain yes yeah what are things we can do and how we can do what are the things are there okay today we are only talking about crickets okay today we can talk about basketball when you talk about us um it's much more of basketball over there when you talk about uk and it's much more football over there okay so again it varies from region to region and basically and if we know how the dynamics of the sports are then we can go really good how to do all those things yeah and so this is this is excel sheet okay i have one doubt from my side uh yeah one specific doubt for example let's say we are working on uh let's say ipl data or or or any kind of let's say icc world cup kind of prediction kind of thing how do we get the fitness of the players because that's very uh i i don't think for a outsider is it very easy to get the data what what is the fitness level of the candidates what's your stake on that no no it's quite difficult it's too confidential data is about the body only so i guess no you can't get data but again if you're getting an opportunity to know about the fitness also it will going to add values in your model okay if if someone is not able to run fast okay it is again going to cause an impact not only it may cause something out okay and again and in the fantasy only if they just run out if and that you have not wasn't over there and he will get extra 25 points over there yeah then the batsman again it pays a roll but again it is quite difficult to know the status how to get just to summarize what you told is in case you are working closely with those teams and you are in that sports industry where you have access to the data in that case is you can probably use the fitness level and build some models but who are not from that domain and probably looking out for some college level projects or something like that for them getting those fitness level data and those kind of confidential data is not possible right that's what you get possible and it's quite difficult to get also okay because there is something which test they are doing you should know what is the significance of that and everything it is also one of the important things over here correct very very nice thank you so suppose this is a data of yesterday match only okay should just to interrupt you i have a question where did you download this ipl data from last year uh so there is a question from nitesh yeah this data sorry which one which one i can get no the uh where can you download the ipl data okay then we can download see the guys that this data of today's session every day ipal is uploading in their website okay so difficult that you have to keep going and downloading the data and making an excel sheet yeah otherwise you you have to wait for whole season to over yeah yeah i mean just to add you sritank even i was playing out with some visualizations and ipl t20.com i think that their website and you can easily scrape data from that so it's quite easily available it is in tabular format and you need to know how to scrape it that's it you can easily scrape it i think i answered your question yeah i should thank you can proceed please so this is the date of yesterday match only okay so name versus mumbai and it is quite easy easy to say karen performed very well over there okay suppose if someone is asking you in a mathematical point of view how to say that okay we can say yeah usually we saw the match he scored a lot so what we can do over here see i have made okay there are batting csk mumbai falling as well and escape boiling as well okay so what you can do you can do over sg score over here okay a simple way to do that okay you are standing standardizing everything over here so happens what but then you can see four was scored by the raj which is scored fifty paired by early then two then seventy two twenty two but if you're only considering about the run over here we are missing lots of thing over here mainly strike rate you can see this point over here okay so this is how it works okay so you can build up simple this and you can use standardized formula from this excel sheet and you can build okay if if you think that yes whose are not performing well you can go see it's negative mark over here okay and if you can give a weightage over here you see he scored 760s you you don't have to give weightage also if you want otherwise you can give it is also okay it's a simple way to do that and you can see over here his score was total 5.7 when we are comparing with the whole team that's why he performed very well in the batting fine so because i have taken into consideration over his track date also okay if you are removing the striker from this picture again the score will change okay you will see that he is performing well in the same way but when we talk about t20 all those things comes in the picture and it does matter it's a simple way to do and you can visualize also okay you can go like this and you can see you can plot simple bar graph okay you can see which is going up which is going down you can place name over here and you can see it's a simple way to see which what is the impact of a player similarly if that here is balling also manually you can go over here and you can check when only one over he gave and he took one wicket also okay his score was this much you can add all those things you can do submission of all those things and you can come over here and you can say yeah it was a i think much more impact over here similarly when you talk about the batting of mumbai indians you can see current polar okay it is a way it is this is the whole data when we are talking when we are talking about the same team but you can do together also you can take hold all the data wire you can put my down you can take overall performance of the team and then the clear picture will come this is the individual we are talking about then you can talk about this way you can copy paste it over here and you can see suppose if i am doing all these things right now see no we can't do right now we have to compare with the mean of everyone we have to again add all those things otherwise if you want i will show you how to do that also fine you can comment on that uh similarly on that basis i did for this india and england i guess it was match okay for the all the batsmen and everything how they performed and everything together questioning okay second in everything and to see overall impact of the match so this is how you can with the simple approach no models nothing you just you have to do x minus x bar upon sigma all those things you can do okay it's not so difficult okay and it's so pretty easy and you can do and you can cut it in a good way whether you want to plot in excel or tableau it's again depend upon you find any no curious there is a question from uh raj paul i don't know her name but yeah she's asking how do we determine the weightage to be assigned on each feature yeah yeah so basically in these things we are not giving any weightage it's it's we are differentiating from the mean okay that is one thing this is not a weightage based things if you want to withdraw weightage three points to the sixth three times if someone is scoring eight sixes the score will go 24 similarly we can go for four days okay we can give weightage but again this all procedure has to be followed for each and everybody in each and every teenage and every team in the season okay and it is sometimes quite difficult sometimes it can it can be confusing also but however you can go for simple this approach it's just comparing with the mean of the team you can see how much they are varying from the mean and you can simply analyze the person how much they are impacting over here fine yeah yeah you can proceed yes yeah so oh yeah although you said main knowledge domain knowledge placing each and everything okay fitness also suppose over here in the boiling i have given a runs minus over here you can see minus one why because the boilers if they are giving much more room much more run over here that's why minus minus weightage the score is much scoring is too much similarly you have to understand all those things okay if the economy rate is too high again it will go for the minus only this is how it works if the person is taking one wicket and giving a two runs again it is not good for the team and all these things analysis you can do okay before the uh this is a historical data again when we are talking about and all these things analysis you can do and you can later on select a player fine [Music] this is how it works it's a simple way for analyzing the impact of the player although you can do with the help of linear regression also you have to provide the features and you can go for the weightage and again with the weightage it will come and you can see which parameter is having which much weightage and depending upon that you can give a basic evidence based approach regarding the weightage system fine now i will be sharing my screen i uh one second and in case you need a one minute break guys just be back in one minute in the meanwhile sri tank will share gaurav has asked so if i apply for a data scientist position in a finance company should i have the knowledge of finance also not not really not really gaurav very off topic but yeah not really you don't have to have a knowledge on finance because i myself having zero experience on finance uh i was able to switch not not a problem i mean yeah why uh domain knowledge plays a vital role but uh mean it's not the case all the time you can easily grasp domain knowledge within a couple of weeks in case you are very good at data science or related topics you can you can easily grasp it but yeah you have to devote a lot of time in doing research so thank uh can you just hide your uh yeah in the bottom here that's it so my screen is coming this one yeah it's visible yes yes okay this is coming up prediction form yeah yeah prediction for means so so fine guys i with the data of my last year only last year whatever the data was basically so i built a model with help of linear regression only and with that model then again i deployed it and this is how it is so you can name players okay i will go go with the code walkthrough also okay okay i will let you know how it is working and just for the over overview i'm just saying this suppose you have written a player name like this okay and his he played 25 matches of which winning he played he was not out five times his average was 68 uh highest score was 112 ball faced 440 striker it was 126 he scored 100 850 is with how many fours 24 and six tens okay and you when you're going to submit it you can see over here it's estimated then will be of overall season will be 734 okay if is if he is playing in like this way okay if you want to predict fine this is how it comes and you can change it also suppose if you want to change the player name also okay and he played some what around some other matches object 25 matches he played his average was you can say 90 and he played much more balls also 560 and secure check it was also better you can say 34 134 you can say 151 okay he scored two hundreds and and okay and 1566 let's see how it comes let's see let's go change this to 967. this is how it works when we talk about production okay in the fantasily okay you can estimate okay again you can you can also do it divide by the total matches and you can come across the individual imagining okay how many is how many then see scoring per match also you can go for that okay this i want to show you i will again show you one more form right now only okay one second okay yeah so this is this another i built for individual okay it changed yeah so this is another episode okay for individual matches okay if you want to check for the individual matches so in that individual matches you're not going to take mac innings over there okay suppose you put without fully again he played around his average was 90 you can see he played in a match around 25 balls okay and his stack it was around you can say 151 again and he hit suppose you're guessing only and only six is five three he is scoring 70 runs in 25 balls okay so this is how it works and when we talk about a prediction in the perspective of fantasy league okay how it works and this i did for you guys only how it is working okay now next i will go through the code work through okay so i think i have a question i have a question i'm just giving you a scenario and let me know if it is possible or not if it is possible we can probably have another session on it okay yeah uh for example uh right now uh you know about fantasy leagues right like applications are there uh i'm not pretty sure dream 11 or some other applications so is it possible based on the previous data and based on the uh till now let's say today's match yesterday's match was on csk versus mumbai let's say we have till yesterday is it possible that let's say today's match i'm not sure today's match is with which and which team let's say for example kolkata vs rajasthan possible to predict a dream 11 team is it is it is it possible like how how difficult is it basically we can go with the two approach over here one was the impact okay because when we want to build a model we need lots of data yeah that is one thing and with this with this season data if you want to predict anything in the future it is quite difficult and it would be biased enough okay that is one thing although we can go for with this approach okay impact okay we can select on the basis of this approach also and yeah we can go with the help of that also we can predict run over here okay as i showed you right now we can go for the prediction of the run source so depending upon that we can select our team and it also depends upon the pitch all those things we can take into consideration all those factors if you want i can build all those things and i will create one data and i will show how easy or difficult it is because i have to work on that aspect also yeah i understand yeah but this is one you're telling that it's possible right yeah it's possible but again the accuracy is one thing over here yeah suppose tomorrow you and me are saying this play will going to play and if the fitness is not good over there and somehow some mishap happened over there then it is quite difficult yeah so yeah it's the data is there with plus minus error okay yeah i mean i understand that kind of things will happen but at least even if with less accuracy but if it is possible then i think people will love to see that application so uh if let's say today i will click on today's match and we'll see show me the dream 11 and we will probably based on our internal calculation we'll show you that these are the list of you know players which people can choose i think that will be amazing i mean yeah i mean i'm not pressurizing you but in case it's possible we can have a session on that yeah we can discuss on that and depending upon that we can go forward with that also nice okay yeah yeah so this is basically your data i took from ipl only okay and this is simple things so importing the necessary libraries over here for data manipulation and some visualization which i didn't do over here okay this okay with help of pandas profiling you can get each and everything about the eda i guess okay pandas profiling will show you in the one line what is the missing value what is the correlation i could listen you can see with something okay so you can again go analyze the eda over here how much missing values are there okay it's a good library to work on the eda along with that one more which you guys can go and work on that also you can go for the interaction you can see you can go for the correlation okay all types of pearson sperm and all those things you can see over here again it depends upon the some sort of knowledge also so this is the data we are talking about okay of the last league i guess okay uh position of the player on the league okay david warner he's got two many rants 692 the average of this much and everything so this was the data okay when we are talking about the data this is the last rows okay and you can see some missing values are there okay so this all things were there so how it worked later on so so you have to replace okay these are the boilers only they will not going to have any sort of average you can replace with one okay it doesn't matter anyhow okay with the okay or you can drop it also again it depends upon you how you go forward with that when you're talking about the batsman over here okay then these are the players okay also plays an important role okay if you're changing the name also the scoring will change okay so you have to create a dummy data again there will be lots of dummy data there because there are 100 players and again you have to drop one from here okay so these are the demo data i have created okay you can see and i'm dropping you rising from here okay and then i dropped players okay because we are having over here you can see one zero all this value fine and then you can see over here lots of columns have appeared 113 because of this one only players okay and this is how it works about the max you can take you have to drop wrench over here okay and not out also okay all those things i have dropped it from here then we are creating a say we saw on when we are putting in and test okay and then we created a linear regression and all those things okay we are doing cross validation over here you see which was much more good and we created one function where the all the features will pass and it will give the return okay so as i showed you over there also i can show you over here also how it works see same thing if i change the name of a player okay if you want suppose i will if you want to kl rahul okay if you want kl rahul over here to see is there any role of those data over here see there's a variation of 11 run over here so that's why we took all the player name over here because it has an impact okay you can't simply drop the player name similarly you can select any baller and you can see how it works suppose if you want to check of andre russell same thing same you are not changing anything if you want to change it you can change it again it varies okay because david warner was performing well as we saw david warner was performing well that's why maybe the scoring was much more however radically was again performing well all those things comes in the picture when we are talking about the code so this is this was a simple code i did okay nothing so sophisticated i didn't play any other model because of the making it simply otherwise we can go for others mathematical approach also over here then we have to dump the model with article and with json we can you can create those model deployment over there and guys it's pretty simple it's simple coding only there is a question that is it possible to predict to win based on batsman and baller uh boilers form and fitness before starting match because i think many people are you know very inclined towards dream 11 and all this application so they are asking about this question is it possible so basically what we can do we are we are able to know about a form we are seeing that there how they are performing the match we can we can follow up those things and we can create one data from that and we can give he is performing well or not we can give a yes and no kind of a thing okay and regarding the fitness again the confidentiality of the although you can say by looking at the match how they are how fit they are okay you can assume but you can't be so perfect over there to say the person is not so fit of it okay yeah yeah yeah so that is one thing but again form you are talking about okay again it depends upon your constitution too many things last year data fitness and that's in the perspective not only bodily mental also and how is how much he is training is there any training load or not again there is one training load okay that is another different topic in the sports domain okay so all the things we can measure it and we can say yeah it's performing good or bad again depends yeah yeah i think the prediction rate which comes when the match starts uh that prediction rate they might be taking fitness into picture for sure right uh i don't think so because see again for the stakeholders players are much more important than fitness yeah yeah yeah if suppose if if you can guess any player okay if they're not so fit but again the name of that player is going to cause an impact to the position plus it does play a role but not in the point of view with the stakeholders when we're talking about because the name is too much of the player and if he's playing the opponent will be like yeah that that batsman will be coming to play or that baller will be going to pole and it is going to cause some trouble and all those things so that's why fitness is the role over here but again it is quite difficult to get a fitness data over here we are having all this data but how much they are scoring and everything and we can predict the run in a similar way we can go for the similar way we can go for the production of wickets also we can just change the features over here and we have to just do a classification model over there no no yeah same model not classification linear regression only okay and with the plane and win and loss when you're talking about specification model we can do yeah yeah yeah and moreover i think fitness i don't think any team will be open to share details about the fitness because yeah because even on espn cricket for any kind of websites before the toss we are not sure who is going to be the playing 11 so we don't know if somebody is injured or something right yeah yeah they keep it confidential yeah i understand that yeah along with that when we are talking about all these things form so we have to take into consideration about the wagon wheel okay suppose uh what is the pitch first thing is that and wagon we love that which is the favorite area of scoring and all those things comes in the picture when we take and if we know about all those things if the ballers we know if someone is that type of bother some spinners are there which is much much more playing then we can say we can assume how [Music] to be selected all those things we can do but again we have put up all the data about that which uh regarding dream 11 and all those things although if you want to predict run and all those things we can do with a simple approach okay although we can make it sophisticated also but it's a simple way and basically i believe this i i i was reading one book okay regarding baseball and i think and i came across how they are doing all those things and you can go forward with this also yeah when we're talking about the sports analytics over here so i would say that it's not about only prediction it's about much more than much more than that okay uh much more things are involved emotions are involved okay right now with the covet situation the stadiums are vacant but if it's but not that the everyone will be going to the stadium okay and the economy is over there all those things comes in the picture okay so that is the thing involving when we're talking with sports and yeah i'm done with all those things yeah yeah yeah thank you tang let's keep the call open for two more minutes in case anybody has any questions uh yeah in the meanwhile we'll also share you this uh regression code i mean it was a very basic idea about what can be done but yeah there are a huge amount of use cases that's possible so i'll try to explore this with sritank and see if it is possible to have a session in the near future but in case you have any questions related to the session or off topic anything else if you want any specific webinars uh you you can just let me know i'll keep the call open for two more minutes guys and then we'll probably wind up so next session is going to be again uh every sunday saturday uh sorry and every sunday 7 p.m i'm going to schedule classes uh webinars workshops and all these things uh for the coming two classes it's going to be on google analytics and i'm planning on a day session on 16th of may uh apart from that i also have some huge announcement i'm not pretty sure if it is going to happen or not but i am planning for a live workshop which is which is going to be over uh four weeks so four weeks four weekend classes live classes on specific topics uh i'll keep you notified through the community posts or through whatsapp groups so i will you will know about it not an issue any question guys if not we'll probably wind up thank you so much for your time thank you shatan for joining it was really nice actually at least for me because i i do care about people but yeah for me it was really nice i'm sure for some of them it was little bit difficult to grasp but yeah it can be consumed later on once you are good with the basics and uh sports analytics is a huge domain uh at least for me it was untouched but i will like to explore more into this it's very fascinating because even i i am very much fascinated into dream 11 and all those things so i'll try to see what can be done a lot of people talk tell me that hey what a data centers why don't you analyze these kind of data why don't you analyze stock markets why don't you analyze cryptocurrencies and i was like i really don't have any idea about that so i will start exploring these kind of things i think there is a latest video on my channel guys if you haven't seen i have done a foreign exchange forecasting that's going to be very helpful for people who are not in india abroad because at least for me when i send back money home i usually look at the exchange rates every day i look at what's the rate today and based on that i transfer money to indian account so uh if you haven't seen that video probably go to my channel the top video you will be able to see it and yeah i'll come up with a lot of interesting topics and lot of interesting webinars as well i don't see any questions here somebody is asking statistics statistics sessions gopi uh many people have asked me about statistics uh i think statistics is very broad so i'm not pretty sure if i can cover it but we can we'll see i'll see if i can uh have a four weeks web workshop or something like that i'll probably plan it very soon quite possible but i'm not pretty sure if you can guarantee me some 20 30 viewers i can do it gopi okay uh anybody has any doubts shaytan's profile my profile are already part of the description you can reach out to us in case you need anything related to sports analytics uh anything related to data science machine learning any kind of things not related to sports area you can reach out to me i would not recommend you to bug shwetank on linkedin uh thank you all for joining the session guys i i will also have few more sessions lined up on various topics so please subscribe the channel and keep me motivated do share my videos as much as possible and like the videos and thank you everyone good night thank you so much thank you so much for joining me thank you so much yeah okay thank you guys thank you good night

Channel: Satyajit Pattnaik

Views: 2,726

Keywords: satyajit pattnaik, python projects, time series, forecasting, data science, Python, python tutorial, learn python programming, python for beginners, sports analytics, IPL, prediction, charts, dashboard, cricket, new python project, basketball analytics, sports jobs, sports projects, nba, data science sports, sports analytics careers, analytics, football, cricket sports analytics, data science project, Everything about Sports Analytics, #sports, #cricket, webinar, streamlit, Swetank Pathak

Id: 2EBDrDzCOTI

Length: 63min 7sec (3787 seconds)

Published: Sun May 02 2021