Hello all my name is krishnak and welcome to my youtube channel so guys today in this particular video we are going to discuss about some of the data science myths versus reality and i'll try to explain in depth the total number of points that i have actually selected are 9 .
So nine myths was it reality i'm going to discuss with you if you are new to this channel please do subscribe the channel press the bell notification icon because in my subscription i can i can see that 40 of you have not subscribed so please make sure that you subscribe because i'm going to come up with some .
Amazing content so let's move ahead so let's go to the first myth you need to have a phd or master degree to get into data science so this is the common myth that people have you know you need to get master's degree phd degree what i'd suggest that is i've seen .
People you know if you are lucky enough right sometimes you need not even need a bachelor degree okay but i'm just keeping this point with respect to phd or masters because there are many questions that have been asked by so many people krish do we require to go with masters to get into data science should i do phd .
To get into data science it is good to have guys but it is not compulsory so this is the first myth please make sure that it is not compulsory to have masters of phd it is good to have right because if you are investing your time in master's degree or full-time .
Master degrees or phd degrees on that specific subject it is good but it is not compulsory it is not possible by everyone to go to masters of phd because there is a lot of expenditure involved i wanted to do my master's degree but due to financial condition i could not .
Do it so i went and took my own path to become data scientist and i'm happy about it right so it is not compulsory to have a master's or phd degree so this was the first myth coming to the second myth guys you need a data science certification to become data scientist .
This is the next thing again certification is not at all required it is again good to have but again understand the the things that we do in the real world use cases is completely different from the things that we learn right so data science certification is .
Not necessary or it is not compulsory that is the point that i want to put i know some of you people may not agree with this but this is from my experience guys and as you all know i try to share my experience with you all i again do not have any data science .
Certification and i have been working in the data science industry from past six years so even when we see i've also seen interviews you know where people have written that we have done this this this certification if you are writing those certification that .
Basically means that we can ask any questions related to that certification right but when we take that interview the person is not able to tell much right instead data science knowledge is important domain knowledge is important how you actually work that is important .
More important than the certification again it is good to have a person from a non-technical background can say that is i've done data science certification since i was actually working in some other domain right so because of that i had to do this particular certification .
So that i portray my knowledge to you all right like this kind of conditions can be put but again it is not compulsory data science knowledge is more important the way that you solve use cases is more important this is the point second that is the second myth and i have explained you .
With respect to reality also coming to the third point your previous work experience is not important this is the myth okay your previous work experience is not important this is the myth that many people have but trust me .
Data science is just not about one type of work your previous work experience will definitely be useful i have seen people who are more than 15 plus years experienced on some specific domain like finance that it be marketing domain let it be retail domain .
They have made a successful transition towards data science with a very very good position right with a higher position and whatever experience they have gained in their previous venture like 15 to 20 years whatever they have actually worked they have used those experience in .
Successfully executing data science project this is pretty much important for all of you guys whatever you do in life is important right because in data science you have to do many things right you have to come up with what data is important .
You have to think about and again i'm telling you domain knowledge is much more important because to initiate that project if we do not have a product owner or a domain experty definitely that project cannot be initiated this is from my personal experience that i've .
Been saying because two of my previous company projects got scrapped because we did not had some good amount of domain knowledge when we did not had a good amount of domain knowledge we were not able to decide which are the important data that we can use in order to execute those data science .
Project so this is the third myth versus reality that i have actually explained right coming to the fourth myth you need to be from computer science stats or programming background now this is the another myth that i have been heard hearing from most of my subscribers for many people when they ask their common queries .
Right and again guys from my experience i've seen that people from different different background have excelled in this data science industry okay it you need not be exactly for computer science background or status background or programming background you can be from any any background as .
Such and as always i explain in my each and every video that data science is a technique that can be applied in any domain so if you are from a marketing domain if you have some different kind of degrees if you are from different non-programming background .
Then also you can excel because learning programming nowadays has become easy and when we see with respect to data science the common programming frameworks that we use are like python or r and trust me just by using google you can do a lot of things i've seen from my senior they're not .
From programming background but just by googling they're able to get some amazing codes and they are implementing it right so this is the fourth myth versus reality that i've explained you need not be from you need not be from computer science stats or programming background but the .
Myth says that you need to be from computer science or stats or programming background so please make sure that it is not compulsory okay coming to the fifth data science is all about model building this is the myth okay again many people when we talk about .
Data science to them they think that okay we are using some machine learning models we are training the models we are building the model that's it most of them still have that specific thought this is the myth guys but in the reality whenever you are solving some real world .
Use cases data science is just not about model building there are so many other tasks that we do right discussion with the product owner requirement gathering you know understanding that which data is important right then following the life cycle the .
Life cycle of a data science project uh storing the data efficiently in some uh databases then in the life cycle we do things like feature engineering feature selection right then over there we have something called as model creation hyper parameter tuning then deployment .
Of that specific model in clouds i'll say that model building is a very very small part in that hardly i can say that less than 10 percent of the overall project because there's so many things that we do apart from that model maintaining model monitoring model .
Retraining approach everything there's so many things apart from that guys so for the people who think that after you become a data scientist or once you get the job it's you just have to go and apply some machine learning algorithm or deep learning algorithm build a model .
No that is not what it is so this was the fifth myth that i really want to discuss with all of you data science is all about model building right now coming to the next myth kaggle or hackathon and real world projects are same this is again a separate myth that many .
People are actually having once they solve some kaggle projects they they think that okay or one they solve some any hackathon project get some good rank the thing that yes they can get into any data science industry and they can world work in the real world project itself again please remove this myth .
Kaggle competition is just to make things clear to you so that you are very good at all the life cycle of a data science project right again in kaggle whenever you're solving competitions model deployment is not there at all you know model hyper parameter tuning is not there at all .
Because the new data that we will be getting right it will be changing frequently with respect to time right each and every time the data will be changing at that time what is the approach that you have to take the retraining approach the model optimization approach the model hyper parameter tuning approach .
Right and this all are actually created as an end-to-end pipes pipeline right in kaggle or hackathon what we do we just do all the feature engineering feature selection we actually create the model and then with respect to attached data we find out the output and submit it in the kaggle or hackathon .
Competition websites and then we see that okay this is good to have right again i'm telling you guys whatever points that i'm mentioning over here as a myth it is good to have but it is not compulsory but this does not decide that how you may be performing in the real world .
Project okay you may you may be practicing for four to five years you can get a better rank in kaggle or hackathon but again working in a data science project is completely different right so this was the sixth myth that i really wanted to discuss please do not consider kaggle hackathon .
And railroad projects are same coming to the seventh myth most of the time of the project is invested in model building okay so many people also think that the huge amount of time from the whole data science project is actually involved in model building right so it is not .
Like that if i talk if i combine future engineering and future selection that alone takes around 30 percentage of the time of the overall data science project right then discussing with the domain expert is requirement gathering and data collection strategy along with storing that data somewhere takes .
Somewhere around 30 to 40 percentage of time right so the most amount of time will definitely be taking in the first stage and the second stage the second stage what i'm doing i'm trying to combine uh feature engineering and feature selection and first stage is about data gathering requirement .
Gatherings and understanding which data are important and storing it into a big data hadoop or amazing databases that we actually have so this is the seventh myth that many people have since again if you are planning to become data scientist you need to know all these things and these all things .
I'm actually telling you from my experience coming to the eighth myth you need to be very good at coding to get into data science again this is a big myth and many people are worried because of this every day i get some calls every day i get some message krish .
In linkedin also i get some message krish i i really don't have that much good coding knowledge can i become a data scientist i have worked in this domain for so many years but i don't have good coding knowledge i'm not good at coding i'm not an expert in coding guys understand coding is pretty simple .
Nowadays to learn google is there who is able to remember all the syntaxes if you just write how to convert a sentence into lower sentence or lower lowercase words you'll just go and google with the help of python java c plus plus c sharp any programming language you'll get that code over there you just .
Have to go and execute it google is there youtube tutorials are there and there's so many things each and every step google is there nowadays if you don't be good at coding it is fine google is there you can actually search everything over there the only thing that you have to be good at is logic .
Building i know where many people have seen they're good at logic building they know that they tell me that okay we have to do these things in the next step in my company also i've seen many people who are not good at coding but their logic building is very very good they know what to do next with respect .
To any kind of data science project so this is pretty much important guys suppose if you're not good at coding don't think that you cannot get into data science you need to be good at logic building and that same thing if you really want to implement into coding just search in the google that's it .
So this was the eighth myth that i really want to discuss and again myth versus reality i've actually explained you now coming to the final point freshers can get ai job okay fresher sorry it should be freshers can't get ai jobs this is the myth that again many people are having i have .
Seen tons and tons of freshers who are smart people who have played with the resume who have done different kind of work who have actually put those information in the resume they're able to get internships they were able to get full-time junior data scientist job in front of my .
Eyes guys they had followed my channel they have followed whatever i have actually said they have made some amazing github profiles they have made some amazing end-to-end projects they have put all the information they have written blogs they have showcased that particular .
Information they have got successfully internship not only passed out students even students who are there in the fourth year they were also able to get it nowadays to get jobs you need to be very very smart there are lot of competitions you need to be the one man .
Out like how can you get the job right and many people think that no freshers cannot get everybody because recently i conducted a live session everybody's saying chris everybody is asking two to three years of experience but when i see job profiles over there also internship job profiles are present .
In linkedin i i do search every day i share within my telegram channel and groups or other groups that i've actually created i share with people over there whenever i get time and if i just go and search in the linkedin then also i'll be getting internships opportunities over there with respect to freshers junior .
Data scientists associate data scientists and many things right so it is very very important guys it is up to you you need the job you have to take that one additional step you have to do some little bit more hard work and find out the crust mechanism to getting jobs and they are many many jobs guys there .
Are many many huge number of freshers jobs right now the covet because of the covet situation now the market is opening and i can see people are getting jobs not only me guys i've seen my subscribers i've seen my juniors i've seen different people so freshers can .
Also get ai jobs please remove this myth because every time i talk to a fresher he says that only krish we are getting two to three years call but we don't have any experience just try to search the right kind of jobs in the market you will be able to get it .
And there are some amazing platforms if you just go and search for linkedin continuously work on your resume with your github profile for continuously one month i think you'll be able to get it so these were the nine myths versus reality that i've actually explained again my final suggestion will .
Be that if you're planning to get into data science industry please learn continuously please please that is a simple request to you all guys over here in data science industry it's all about logic building it's all about understanding the data it's all about understanding your domain creating the web application creating .
The data science application will be very very easy when you are very good at that specific things okay so finally guys uh i'd also like to ask some questions that please do comment down in this particular video if my youtube channel that have helped you all to learn something with .
Respect to data science or to achieve the goal that you had actually planned with respect to data science or machine learning engineers or deep learning that would be very very grateful for me i really want to take this kind of feedback apart from that guys please do subscribe my ai tech news channel .
That is krishna vlogging channel the link of the youtube channel will be given in the description so i hope you like this particular video please do subscribe the channel and i'll see you all in the next video have a great day ahead thank you one and all bye bye