Thanks for the 8-min summary! The part about pairing a data scientist with internal people in an organization, and why that's important is spot on. I know an org that has exactly this problem. -rolf
I graduated from both Electrical engineering and information system science. I started my career as a programmer analyst, system analyst, and business intelligent analyst. Then later on, I became a data analyst where I used mathematics such as linear regression, auto-correlation, ARIMA, Hyper-geometric, and Cochran's Theorem to analyze and present data in a meaningful way. So Mike, am I qualified to be a data scientist?
well said Sir. It needs more than programming skills or any experience related to that but a lot of skills consolidated to visualize a problem and apply those DATA skills to understand and Solve a problem not just for a business but close to a lot of Data.
Seems from Mike's definition that anyone who is (and enjoys being) super analytical can do the job? People who studied humanities - philosophy, economics, english, history, etc...are used to breaking down any type of knowledge to its atomic parts and then putting it back again to form a larger picture and in the process making note of patterns...
@sel Very much so. One of the things that Mike points out is that a data scientist forms a hypothesis about what the data might tell us. That kind of insight can definitely come from people in non-tech fields. There will be more to learn, but the basic tech for data analysis has never been easier. Don't let anyone discourage you from considering this field (and that's from someone with a CS degree!). Good luck to you! -r
3:03 In spite of that, I wonder why it would seem that Netflix's recommendation engine is worse now than when it began? I think it relates to the system getting smarter at correlation but with limited ability to discern causation. For me, it's probably coincidental if the movie soundtrack was done by the same person in my top movie picks, but the algorithm has already noted a strong correlation that I prefer movies with certain musical scores (a made up example). It seems like there's room for a lot of this sort of error in the field. I've noticed Netflix never gives a quick validation: "ok you gave this a thumbs up, was it mostly because of A, B, C, Other/None of the above." That would probably help. I remember the moment where Netflix decided to go from: "here are the ratings of other users" to "here's what we think you'll rate this movie," which was always WAY off. So, I no longer had any help finding things I'd like. You liked this one movie that most people disliked? It must be because you love bad movies. Here are all the worst movies. You're welcome. That was a terrible change. It's been that way since. I think it's easier for a human to predict and navigate a simple algorithm, than it is for a complex algorithm to predict human behavior/choices/preferences. So, the use of such algorithms makes sense for predicting things which do not depend on preference, but are instead complex interactions which seem random, such as modeling weather. It can absolutely be a hindrance elsewhere if forced, such as in the Netflix example.
great to know that, honestly i know i make somthing. that i always did when i study on my own sf study. but because out of budget and much better resources sometimes i feel that i need someone who ca help me and or what ever it is. in short because i cant do anything for now i hope they develope it for us for the goodness of our system who ever found it. I hank you..
Hi Mike and others! Right now im finishing my PhD degree in ecology, but i graduated in physics. Right now i have good handling in programming, statistics and others. Last months i have been researching and reading a lot about data mining and data science in general, i can say that i have an adequate management about algorithms and usages. But given that i have been working ONLY in the academy i have no idea how to start working in data mining. I dont really know if some enterprise would hire me if i show my scientific publications. Can you give me an advice about how to start in the area? How i present myself? Thank you all
+Nassim Dhaher I think you need to add in domain knowledge.. You can't have someone that run statistics or algorithms if they don't understand the data... Take biology for example.. you don't take a random person with machine learning, statistical analysis and utilizing databases and start mapping the human genome... Unfortunately this video is a little dated.. The most recent iteration of the Venn diagram he's shown has 3 circles instead of two.. domain knowledge, stats and programming..
I don't think people use algorithms to confirm a hypothesis. It should go like this. The real world poses a problem, come up with a solution using an algorithm, tests the soluntion in the real world then conclude.
Thank you Mike. I am 19 years old and I want to discover new trends in my online community. In my high school, I was among the top students in math. How should I start now to be a very successful data scientist in the next 5-10 years?
This is my personal path to become data scientist, really weird actually, but here it is: Highschool (computer specialization)+ Actuarial Science Bachelor + Hacker Hobby + Data Engineering Master + Big Data PHD. Well, i am trying to take this path, now i was accepted for the master, so hopefully i will become a data scientist soon. Twitter Firehose here i come :).
How much time do you think it would take me (programmer) to be a data scientist and is it necessary that a data scientist should learn Hadoop programming.
I am a recent Science graduate with a major in Physics. I have no relevant work experience. How does someone like me break into the field and become a Data Scientist? Any help would be appreciated!
why the hell are u guyz bothered about his shirt !! r u guyz looking for a perfect person who wear neat and tidy dress or bothered about the content which is he talking about! crazy people
Hi Mike, Thanks for your such an excellent explanation. But I am sorry to say that I am barely understand what you said. As your voice is echo. Can you please again record and upload .
I am a theoretical quantum phycisist, finishing my phd in approx. a year. So I basically have highly advanced knowledge of mathematics, especially linear algebra. I am not really a programmer, but I have had to use some numerical/coding tools such as matlab. Which big data companies would be interested in hiring someone with my skills? I still have to decide if I want to go into finance or somewhere else after my doctoral studies. However, I recently heard more and more about the growth of big data science,
Right now you have a chance of getting hired as an entry level Consultant but not as a Data Scientist because your background is very theoretical. Computer Science, data mining, and IT skills are very important to companies looking for Data Scientists. Also, being good at math does not necessarily mean being good at data analysis and statistics (I know that too well). You should follow a postgraduate degree to specialise yourself at something more applied and get some strong IT skills, then your profile will be unbeatable.
Raymond Martin To reply to your previous comment (which is now unrepliable), these days most people perceive "Data Scientists" as people who work exclusively in the industry. Most Data Scientists have a PHD and therefore a background in research but they're not called "Data Scientists" by anybody until they join the industry. If you have all the skills of a Data Scientist and work on Bioinformatics then you're a Bioinformatician, simple as that. I don't know what your book says but from my personal experience I would say that the industry looks for two profiles of Data Scientists: (1) a specialised Data Analyst with advanced IT/programming skills and (2) a specialised Computer Scientist with experience on Hadoop and Big Data. So I agree that someone with a Computer Science background can easily become a data scientist (the second category). Also, people from Computer Science may have different profiles. I've known people from Computer Science who were purely tech freaks and others who were math whizz. Then again you have Theoretical Computer Science which is basically pure mathematics. Yes, Computer Science is at the heart of all data analytical techniques but it all depends on the person and it's hard to make generic rules.
Raymond Martin I don't know what will happen in the next few years but I know what's happening now. Today the competitive Data Scientist profile that the Industry looks for must have mixed skills in Computer Science and Data Analysis, preferably with a quantitative PhD research background. The thing is that those "MSc. Data Science" degrees are very new and not well established in the industry (yet), hence why today's Data Scientists are expected to have a multidisciplinary background of different degrees where a PhD is a huge plus. The profile that I'm describing will be even more competitive in a few years when anybody with some BSc and/or MSc will call themselves "Data Scientists" though the Industry will never favour them.
Raymond Martin Raymond Martin I haven't read books or surveys but I have worked for and gone to many "Data Scientist" interviews with multi-national giants. Data science is a specialty and not a discipline per se, where the postgrad degrees are mostly a mix of quantitative/computational and technical computing skills with a focus on Big Data. The problem with such a "Data Science" postgrad is that it is a purely professional degree and it makes no sense to go into research after that. Generally speaking, a PhD may not be the norm but it's a huge plus for a quantitative post because it shows that you have a background in research. Smaller to average companies don't care about that so much, as you said, a postgrad in data science may suffice. In big companies, however, where competition is much higher, a PhD and background in research combined with quantitative/technical skills will get you more noticed than someone with a "Data Science" postgrad (assuming no previous experience in both cases). I'm not speaking hypothetically, I know this first hand both as an employer and past candidate. Some of the largest Finance, Consulting and IT companies perform automated CV screening by scanning degree names and searching for a PhD. It should be noted that once you reach the technical interview stage, your background doesn't matter anymore, then it all comes down to performing well at the interviewer's tests/questions/problems at hand.
+Benoit Descamps why would someone with a phd in quantum physics want to break into finance or data? Im mainly just curious, they seem like very different skill sets altogether, and it seems impractical to go for a phd in physics if you know that you have other intentions. Although i can understand if you have had a change of heart towards physics
could u please say sth about the business analytics major's placement? because it is not professional as computer science and statistic , nor professional from finance guys.
+ranfeamon You can find a lot of tutorials for the main factors in data science here on UA-cam. There's a lot of math and language arts in it. If you think you're good at visualizing data, as in mathematical problems, and if you're good at reading and adapting to new languages, then this is the job for you. If you don't find yourself good at these starting factors, then this is not the best job for you. If you feel eligible for this, then Universities like MIT are great. I graduated MIT with a masters degree in computer science and a bachelors in data science. I earn over 400,000 annually, so it's a great career. There's not a lot of us, which makes it a great choice in career, but it takes A LOT of skills. A lot of domain knowledge and comprehension of code and language variation is an essential characteristic you must have. As in computer language, start off with easy ones like HTML and Python. We use these two most often.
+Golden Hawk * A bachelors degree in Chemical Science, not data science, sorry about that! When I went to MIT there were no Data Science careers. It was quite new back then, but it may be a career choice now, however I'm not sure. I started off as a computer scientist who programmed websites for small companies, but I moved under the title of a "Data Scientist" since I thought It'd be a good adjustment to my job.
I am a Java Programmer and I want to do PhD with Data science as specialization. But its getting difficult for me to find data science specialization. How do I choose this program for my PhD . Do I have to choose data mining and algorithms for data science.?
i'm have 2 yrs exp from electronics background but whenever i went to interview for data analyst they ask me ab8 my work experience in data science i told them i completed data science project during course , i don't know what they really want everyone has different requirement for jobs other than M.L, Python , R , Statistics some of them want SQL, Tableau , Hadoop , AWS , NLP , AI, Deep Learning should I have to learn all of this ..?
I saw Dr. Mark Rappa present at the SAS Analytics conference last week and was very impressed by the program he's established at North Carolina State University. analytics.ncsu.edu/
Ahmad Taha Big Data is just large amounts of data ... that's all Data Science is making hypothesis and using algorithms and Big Data to prove those hypothesis ... Like he said in the video, "Scientists make new discoveries" I hope that answers your question.
Does Bob Marley know you're blocking the door to his office?
+Aaron Wolbach lolol
I was holding the camera..
He Jammin'
Well spotted Rainman! :) You should get yourself to those Vegas Blackjack tables. you don't miss a trick! :)
Why Arjen Robben has been data scientist?
Epic comment,
When i first see him,
I am also think robben
One of the best explanation for a Data Scientist I have ever heard.. Thank you Mike !!
Mike, I must say this is a great explanation for a very complex topic. Thank you for your help.
You have a superior illustration skills, thanks a lot bro.
thanks
Perfect explanation! Bravo
simple and excellent explanation of what is and not a Data scientist. Thanks!
Thanks for the 8-min summary! The part about pairing a data scientist with internal people in an organization, and why that's important is spot on. I know an org that has exactly this problem. -rolf
Very good video, very interesting, thanks you so much from France !
Thanks for giving better clarity on what we are doing and what we want to be.. Thank you.
Mike, this was really helpful. Thanks for sharing!
Interesting I came across with the term Data Scientist when I was learning how to use Hadoop, many thanks for your explanation.
you have explained it well , i am an economics graduate and I have understood properly what data science is
Very informative! Thank you.
I graduated from both Electrical engineering and information system science. I started my career as a programmer analyst, system analyst, and business intelligent analyst. Then later on, I became a data analyst where I used mathematics such as linear regression, auto-correlation, ARIMA, Hyper-geometric, and Cochran's Theorem to analyze and present data in a meaningful way. So Mike, am I qualified to be a data scientist?
well said Sir. It needs more than programming skills or any experience related to that but a lot of skills consolidated to visualize a problem and apply those DATA skills to understand and Solve a problem not just for a business but close to a lot of Data.
Thank you Mike, great intro.
Seems from Mike's definition that anyone who is (and enjoys being) super analytical can do the job? People who studied humanities - philosophy, economics, english, history, etc...are used to breaking down any type of knowledge to its atomic parts and then putting it back again to form a larger picture and in the process making note of patterns...
+Zachary Benning they do in higher levels. And we use logic in philosophy to analyze, not feelings.
@sel Very much so. One of the things that Mike points out is that a data scientist forms a hypothesis about what the data might tell us. That kind of insight can definitely come from people in non-tech fields. There will be more to learn, but the basic tech for data analysis has never been easier. Don't let anyone discourage you from considering this field (and that's from someone with a CS degree!). Good luck to you! -r
Thanks, Rolf W!
Mike, thank you for the clear insight.
Excellent and Thank You Mike.
Very Clear Explaination. Great Job and pls continue do so.
Thanks
Very informative video on data scientists
At last...someone who can explain my job in a simple way!
3:03 In spite of that, I wonder why it would seem that Netflix's recommendation engine is worse now than when it began? I think it relates to the system getting smarter at correlation but with limited ability to discern causation. For me, it's probably coincidental if the movie soundtrack was done by the same person in my top movie picks, but the algorithm has already noted a strong correlation that I prefer movies with certain musical scores (a made up example).
It seems like there's room for a lot of this sort of error in the field.
I've noticed Netflix never gives a quick validation: "ok you gave this a thumbs up, was it mostly because of A, B, C, Other/None of the above." That would probably help.
I remember the moment where Netflix decided to go from: "here are the ratings of other users" to "here's what we think you'll rate this movie," which was always WAY off. So, I no longer had any help finding things I'd like. You liked this one movie that most people disliked? It must be because you love bad movies. Here are all the worst movies. You're welcome. That was a terrible change. It's been that way since.
I think it's easier for a human to predict and navigate a simple algorithm, than it is for a complex algorithm to predict human behavior/choices/preferences. So, the use of such algorithms makes sense for predicting things which do not depend on preference, but are instead complex interactions which seem random, such as modeling weather. It can absolutely be a hindrance elsewhere if forced, such as in the Netflix example.
Thank you for this video! You are a very interesting presenter.
Thanks Mike - succinct and helpful to my work!
A good summary of data scientist definition.
that's what i was searching for ; thank you thank you from the deep :)
Thanks ... very good information ...
great to know that, honestly i know i make somthing. that i always did when i study on my own sf study. but because out of budget and much better resources sometimes i feel that i need someone who ca help me and or what ever it is. in short because i cant do anything for now i hope they develope it for us for the goodness of our system who ever found it. I hank you..
Liked it. Thanks for explaining very well.
Thank you for a thorough video, very helpful!
For those taking my programming lessons.
i love u
Mrayis100 "I believe ya [pulls out Tommy Gun] but my tommy gun don't."
Wild Academy hahaha
Good information thank you
Finally, someone who explains things correctly!
Mike,Excellent presentation
Mike... that's a great explanation.
I'm pursuing a marketing career path, but I do love statistics. I might consider becoming a Marketing Data Scientist
Awesome presentation, try to make video on related programming only
so that's like statistics with algorithms which is hard
Hey Mike,
Please also inform us of this Bob Marley that sits behind where you're speaking. Thanks for the great video.
One of the core fundamentals in computer science is developing algorithms.
Hi Mike and others!
Right now im finishing my PhD degree in ecology, but i graduated in physics. Right now i have good handling in programming, statistics and others. Last months i have been researching and reading a lot about data mining and data science in general, i can say that i have an adequate management about algorithms and usages. But given that i have been working ONLY in the academy i have no idea how to start working in data mining. I dont really know if some enterprise would hire me if i show my scientific publications. Can you give me an advice about how to start in the area? How i present myself?
Thank you all
Good explanation
Great Tutor
Thank u sir, its a highly informative video.
Great explanation.
Statistical Analysis + Database data + Mathematical Algorithms = Data Science
+Nassim Dhaher I think you need to add in domain knowledge.. You can't have someone that run statistics or algorithms if they don't understand the data... Take biology for example.. you don't take a random person with machine learning, statistical analysis and utilizing databases and start mapping the human genome... Unfortunately this video is a little dated.. The most recent iteration of the Venn diagram he's shown has 3 circles instead of two.. domain knowledge, stats and programming..
I don't think people use algorithms to confirm a hypothesis.
It should go like this. The real world poses a problem, come up with a solution using an algorithm, tests the soluntion in the real world then conclude.
A master in mathematics is key for this job.
Excellent!
Thanks a lot . Now I'm more motivated to study it
Very clarifying!
awesome
Very good description.
Thanks Mike Gualiteri
Thank you Mike. I am 19 years old and I want to discover new trends in my online community. In my high school, I was among the top students in math. How should I start now to be a very successful data scientist in the next 5-10 years?
this working out?
muy claro y acertado
good video...
Just Excellent >> +10,000,000 :)
It's the dude that knows his way around Ms Excel and ODBC ?
Rastafari Live Up. Enjoy the Bob Marley picture in the back
kansashick must be a running joke for that guy in the office
Thanks a lot.
Thank you.
thanks for clear my doubt bro
What's the deal with Bob Marley room/office Rm 3330?
Hi Mike,
An excellent presentation.
Please can you recommend a good training company for this? Many thanks
Linda
Bob Marley on the wall discovering data science.
This is my personal path to become data scientist, really weird actually, but here it is:
Highschool (computer specialization)+ Actuarial Science Bachelor + Hacker Hobby + Data Engineering Master + Big Data PHD.
Well, i am trying to take this path, now i was accepted for the master, so hopefully i will become a data scientist soon. Twitter Firehose here i come :).
Jah Rastafarai, Bob Marlay famalam big up, Al Gore Riddims man.
Thank you so much for such a precise explanation!
nice info :)
An office for Bob Marley, brilliant :-)
you can find data science tutorials at our channel
How much time do you think it would take me (programmer) to be a data scientist and is it necessary that a data scientist should learn Hadoop programming.
I am a recent Science graduate with a major in Physics. I have no relevant work experience. How does someone like me break into the field and become a Data Scientist? Any help would be appreciated!
why the hell are u guyz bothered about his shirt !! r u guyz looking for a perfect person who wear neat and tidy dress or bothered about the content which is he talking about! crazy people
the man is giving information and STILL they complain. Fuck heads!
quality of shirt = data
bad data = bad algorithms = bad data scientist
bad quality of shirt = bad data
do the math lol.
Askance ugjqqvioii
Good vid
Hi Mike, Thanks for your such an excellent explanation. But I am sorry to say that I am barely understand what you said. As your voice is echo. Can you please again record and upload .
thank you for that
I am a theoretical quantum phycisist, finishing my phd in approx. a year. So I basically have highly advanced knowledge of mathematics, especially linear algebra. I am not really a programmer, but I have had to use some numerical/coding tools such as matlab.
Which big data companies would be interested in hiring someone with my skills?
I still have to decide if I want to go into finance or somewhere else after my doctoral studies. However, I recently heard more and more about the growth of big data science,
Right now you have a chance of getting hired as an entry level Consultant but not as a Data Scientist because your background is very theoretical. Computer Science, data mining, and IT skills are very important to companies looking for Data Scientists. Also, being good at math does not necessarily mean being good at data analysis and statistics (I know that too well). You should follow a postgraduate degree to specialise yourself at something more applied and get some strong IT skills, then your profile will be unbeatable.
Raymond Martin To reply to your previous comment (which is now unrepliable), these days most people perceive "Data Scientists" as people who work exclusively in the industry. Most Data Scientists have a PHD and therefore a background in research but they're not called "Data Scientists" by anybody until they join the industry. If you have all the skills of a Data Scientist and work on Bioinformatics then you're a Bioinformatician, simple as that. I don't know what your book says but from my personal experience I would say that the industry looks for two profiles of Data Scientists: (1) a specialised Data Analyst with advanced IT/programming skills and (2) a specialised Computer Scientist with experience on Hadoop and Big Data. So I agree that someone with a Computer Science background can easily become a data scientist (the second category). Also, people from Computer Science may have different profiles. I've known people from Computer Science who were purely tech freaks and others who were math whizz. Then again you have Theoretical Computer Science which is basically pure mathematics. Yes, Computer Science is at the heart of all data analytical techniques but it all depends on the person and it's hard to make generic rules.
Raymond Martin I don't know what will happen in the next few years but I know what's happening now. Today the competitive Data Scientist profile that the Industry looks for must have mixed skills in Computer Science and Data Analysis, preferably with a quantitative PhD research background. The thing is that those "MSc. Data Science" degrees are very new and not well established in the industry (yet), hence why today's Data Scientists are expected to have a multidisciplinary background of different degrees where a PhD is a huge plus. The profile that I'm describing will be even more competitive in a few years when anybody with some BSc and/or MSc will call themselves "Data Scientists" though the Industry will never favour them.
Raymond Martin Raymond Martin I haven't read books or surveys but I have worked for and gone to many "Data Scientist" interviews with multi-national giants. Data science is a specialty and not a discipline per se, where the postgrad degrees are mostly a mix of quantitative/computational and technical computing skills with a focus on Big Data. The problem with such a "Data Science" postgrad is that it is a purely professional degree and it makes no sense to go into research after that. Generally speaking, a PhD may not be the norm but it's a huge plus for a quantitative post because it shows that you have a background in research. Smaller to average companies don't care about that so much, as you said, a postgrad in data science may suffice. In big companies, however, where competition is much higher, a PhD and background in research combined with quantitative/technical skills will get you more noticed than someone with a "Data Science" postgrad (assuming no previous experience in both cases). I'm not speaking hypothetically, I know this first hand both as an employer and past candidate. Some of the largest Finance, Consulting and IT companies perform automated CV screening by scanning degree names and searching for a PhD. It should be noted that once you reach the technical interview stage, your background doesn't matter anymore, then it all comes down to performing well at the interviewer's tests/questions/problems at hand.
+Benoit Descamps why would someone with a phd in quantum physics want to break into finance or data? Im mainly just curious, they seem like very different skill sets altogether, and it seems impractical to go for a phd in physics if you know that you have other intentions. Although i can understand if you have had a change of heart towards physics
Very informative. Big up to the Bob Marley picture in the background..
could u please say sth about the business analytics major's placement? because it is not professional as computer science and statistic , nor professional from finance guys.
When you buy a new shirt and put it on before ironing...
flopkuff p
But what should I study to being a data scientist
Hi, I want to know the best place to study this for a master degree, and if is self-learning difficulty. thanks
+ranfeamon You can find a lot of tutorials for the main factors in data science here on UA-cam. There's a lot of math and language arts in it. If you think you're good at visualizing data, as in mathematical problems, and if you're good at reading and adapting to new languages, then this is the job for you. If you don't find yourself good at these starting factors, then this is not the best job for you. If you feel eligible for this, then Universities like MIT are great. I graduated MIT with a masters degree in computer science and a bachelors in data science. I earn over 400,000 annually, so it's a great career. There's not a lot of us, which makes it a great choice in career, but it takes A LOT of skills. A lot of domain knowledge and comprehension of code and language variation is an essential characteristic you must have. As in computer language, start off with easy ones like HTML and Python. We use these two most often.
+Golden Hawk * A bachelors degree in Chemical Science, not data science, sorry about that! When I went to MIT there were no Data Science careers. It was quite new back then, but it may be a career choice now, however I'm not sure. I started off as a computer scientist who programmed websites for small companies, but I moved under the title of a "Data Scientist" since I thought It'd be a good adjustment to my job.
I am a Java Programmer and I want to do PhD with Data science as specialization. But its getting difficult for me to find data science specialization. How do I choose this program for my PhD . Do I have to choose data mining and algorithms for data science.?
This dude needs an iron. That shirt is fresh.
So what's the difference between a programmer and data scientist.
thx
i'm have 2 yrs exp from electronics background but whenever i went to interview for data analyst they ask me ab8 my work experience in data science i told them i completed data science project during course , i don't know what they really want
everyone has different requirement for jobs other than M.L, Python , R , Statistics some of them want SQL, Tableau , Hadoop , AWS , NLP , AI, Deep Learning should I have to learn all of this ..?
5:29
Aka Poolka glitch
Hi Mike Gualtieri, do you happen to know a good master program for data scientist in the USA or Canada? Thank you
I saw Dr. Mark Rappa present at the SAS Analytics conference last week and was very impressed by the program he's established at North Carolina State University. analytics.ncsu.edu/
Hopefully I fall in the middle as I have a MIS degree with a masters in ERM. Data Scientist must have those ah-ha moments.
Brand new shirt eh?? Haha
i am currently pursuing my mba in it is there any particular career path you have to follow to become data scientist?
Thank you
the god ida
what is the deference between Data Science and Big data !!!!
Ahmad Taha Big Data is just large amounts of data ... that's all
Data Science is making hypothesis and using algorithms and Big Data to prove those hypothesis ... Like he said in the video, "Scientists make new discoveries"
I hope that answers your question.
Plz tell me
my DREAMS from Somalia