Thanks for watching everyone! Definitely check out the data professor's channel as well! More great visuals of his can be located here: github.com/dataprofessor/infographic/blob/master/06-Different-Data-Science-Roles-Explained.JPG
I really like the fact that the description is basically a transcript! Sometimes you just can't watch the entire thing, and it really helps, This definitely cuts into Ken's watch time, awesome selflessness mate, Thank you
Wow! This video was definitely an up in production. It is really awesome to see you smile when you discuss your favourite topics in DS or f.e. when you did the intro, it really shows that you're passionate about what you do and you enjoy making videos. The nice graphics were a great addition to your video, especially since they really fit well with a topic like this one - showing what role is located where in the project lifecycle. I'm excited to se what you come up with next! :D
Thanks! I love your content; definitely drew some inspiration from your software engineer roles video. Not sure if you would be interested, but I think it would be fun to collaborate on a video!
23k subscribers! You are really blowing up. Not long ago you were under 3k, then 10k, then wow! It's great to see you are growing. Keep up the great content.
Excellent resource for understanding where each role fits in. You're right that there is a lot of flexibility in roles (fair amount of overlap) and I'm sure it is expected that the smaller an area that one role focuses on, the more specialized they will be in that role.
Thank you again for this video, Ken. We can say that I already work as a data analyst, however my tools are restricted to BI and excel. It turns out that I have the business core, lacking programming and math skills. Your videos have been helping me to create a plan in this process to become a data cientist. thank again bro
@@aqibhuda7652 Not the OP, but I'm currently taking this same path. I started with getting a bachelor's degree in mathematics with a minor in statistical modeling, and applied to data internships during my junior year. I've been at my current internship as a data analyst for about a year, and the company I'm working for allows me to collaborate with the data science team a bit to get some experience in that realm. Hope this helps!
Thanks for this video. I reminded me of this book "Making sense of Data" by Glenn J. Myatt and Wayne P. Johnson It's focused mostly on the first four parts from what I saw.
I have a business role in Operations / Planning, but all my major accomplishments were related to collecting, treating and analyzing data in Excel and Power BI. I am currently learning Python, R and statistics to do it more professionally, and in the future I would like to be able to do predictions/ML to bring even more value to the business. I fear that strict data roles may get me too far from the business routine, what could be a problem as I do enjoy living in the supply chain area. I am not sure if being a Data Scientist is my ultimate goal, but I also don't know which roles would be more aligned to my objectives. How do you think Data Analyst/Data Science roles would blend with more business roles?
Thanks for watching Paula! I think a product owner role could be a really good fit for someone with your skillset. There are also some data science roles that are more business facing. These vary by company and seniority, etc. I think there is a huge value in data scientists having more business knowledge though.
@@KenJee_ds Thank you! Your videos are so clear and heartwarming for us trying to get in such a wide area. At the moment, the strategy I have on mind is move to a Data Analyst / Analytics / BI role and see if I really don't like the "strict" data roles or if I'm just fearing the change. After that I think it will be more clear for me if I should go for a senior business role and then apply my data abilities, or to senior data role and help the business as a whole. I think without that experience I will always slip on the "what if" of not knowing enough of the other side.
Between EDA and “model building”, where would you include feature engineering, feature selection , model validation? I ask because I have sadly seen MLE job postings where there is zero “modeling” work. I often find myself in this very situation at my current job where I merely productionize models created by more senior data scientists. So yea I wonder how you think about the different interpretations of these roles in the industry =) .
This is a really good question! I think that feature engineering can come in data cleaning, the beginning of EDA or during the model building process. I don't think it really fits neatly into a single one of the steps. For feature selection and model validation, in most cases, I would put them into the model building faze entirely. I think that your experience really brings up one of the greatest challenges in the data science field: no clear definition of each of these roles. Thanks for watching and I hope this helps!
Thank you Ken, because of your video, im into Data Science world. currently I work as a Data Analyst for about 7 months now. I still find it hard sometimes to learn Data Engineer and DS part. Is it better for me to learn DS part or DE part first? thank you
I think you should work on expanding into the area you are most interested in! I personally would try to explore more of the model building stuff first. That is more important than data collection for a data scientist in my opinion.
@@kristenlobo4161 It's ok, i love to share. like Ken said. There are a lot of type of data analyst /scientist but mainly you do general things like in this video. What I love being a DA is you can breakdown data and gain a lot of new information. For my case, I work mainly with e commerce business and from our DA work we can talk to business owner and discuss an action or two from the data we process. Sometimes hard but I love to learn more about this DS world. You could find more info in kaggle and github and connect Ken Jee in linkedin. Hope this will help
I think that could be a good option depending if you like the cs work! There are definitely parallel skills, so it should be possible to pick up the concepts.
I’ve been kind of afraid to ask this question, but what do you think is the major difference between a data scientist and statistician? Sometimes the line gets very thin between the two
Don't be afraid! Data scientists generally do a bit more with programming and creating tools with their analysis. Statisticians are more focused on creating reports and recommendations with their findings.
Really clear description! Until now i have found the model deployment step really tiring and complex. Is it necessary to master if i want to become a ML Engineer?
Hi Anurag - Thanks for watching! Unfortunately most MLE's need to be able to deploy models. This can be made far easier with tools like Sagemaker though!
You can study Data Science and Artificial Intelligence in my university as a Bachelor and Master - would you consider that a valid option - as Artificial Intelligence seems to be more needed for someone with a PhD, would you therefore recommend starting with studying raw Data Science, Data Engineering or even Machine Learning for a Master? Kind regards~
If your goal is to become a data scientist. All of those majors are viable. They all have core underlying classes in data science. I don't know what country you are from, but I have not heard of a data engineering masters in the US.
I think that either of those options are fine. If you aren't particularly interested in AI, I think it is fine to do any of those other degrees. I personally did my Masters in Computer Science With a Concentration in ML & AI. I don't really use any of the AI stuff that I learned at all though. I would check out this video to better understand the differences between the fields: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-q8cEt8gj3zY.html .Thanks for watching and I hope this helps!
Great content Ken! Do you know ir there’s any DA or DS internship on FANG companies? I could only found a research intern position, and I’m not sure it is trully related to DS.
They should have them, but I don't know where they are posted. I would also check out the virtual internship options on the data professor's channel. Those are great options as well.
So I'm trying to be a data scientist eventually,and I know from experience I don't like software engineering but I do really enjoy the role I had as a data analyst intern.Would you say looking for a job as a data analust is a good way to gzt your foot in the door to work up to data scientist eventually? I also can't afford to do an advanced degree.
Hello sir I know this question is coming out from no where, but I feel you are the right person to ask the question so I'm asking how to select best 'Categorical features' in classification problem? is there any way to find which categorical feature is contributing more in terms of predicting?
If you use a logistic regression you are able to see the coefficients for the categorical variables. If you are using a random forest, you can see the feature importances. I would explore those two types of models. I hope this helps!
Good stuff, man. I'm interning as a Data Engineer this summer, and was also one last summer(i actually really enjoyed it). Mainly focused on ETL's and automation using Python and AWS... Do you have any tips on resources where I can better my SQL knowledge?
Thanks for watching! I think the data engineer role is one of the most important ones. Without good data all models are bad! I don't have any resources off the top of my head, but if I come across one I will let you know!
First of all thx for sharing your knowledge On question, Which is better to deploy a model in your opinion(flask or django) Which is easier to deploy (flask,django)
I think flask is better for both! Django is for more advanced web dev tasks. Generally as data scientists we are making API's so flask is plenty. Thanks for watching so many of my videos recently!
Hey Ken, great video as always. Just a small question, im currently an industrial engineering student (so i have a good background in statistics) and have enrolled in the EDX IBM data science certificate program. I have finished all the courses (excpet for the SQL one which i will start soon) and feel like im fine with the coding and concepts (regressions, clustering, classification, etc. ) but do not know where to go from here since i have only worked on 1 project (the capstone from the course). As a next step i am thinking of learning tableau (for data visualization) and SQL, but apart from that what do you suggest i do ? Thanks
I recommend going to kaggle.com and starting to do some more projects with data there. You can also collect your own data and do projects with that. I think projects are the best way to further your skills and make you more desirable to jobs. I hope this helps!
@@KenJee_ds Hey Ken, sorry to bother again but I think you are one of the few people on RU-vid who interact with their audience and reply to all their questions to truly benefit them, so I have a small question and if you can me a little bit of guidance. I decided to work on a project myself, more specifically related to the movie industry (movies are my passion btw). Considering that the amount of studies done on the correlation between box office, audience ratings, critics ratings, etc. is very huge, I thought of something I have not yet found a study on, and that is the importance of a movie's first trailer towards its overall box office, audience and critic scores, since I think (based of experience of watching trailers) that a movie's first trailer is the most important part of the marketing campaign . The variables I am considering using are views, likes, dislikes of the first trailer (on the movie studio's official RU-vid channel), as well as rotten tomatoes audience and critics score, IMDb score, as well as Metacritic score. However, the main issue is how would I gather the data ? Do you recommend gathering it manually (which will lead to have a small data set and sample size) ? I do not have a lot of experience web scrapping, so if that is the way to go can you please briefly explain to me how to properly use it, since the data I need comes from various sources/websites (RU-vid, IMDb, rotten-tomatoes, etc.) ? One final thing, do you think it is a good project to work on and gain proper experience in the field of movie analytics ? Thank you so much and waiting for your insight.
Hi Ken thanks for making this video. I found this very helpful. I have a question. As I am in my senior year of computer science student, I am not sure whether I should go further into master. Especially, when I look up the job openings, all data scientists are usually required to hold at least a master degree. Is it necessary for a data scientist or relatable positions as you mentioned in this video to have a master or even Ph.D in Data Science field? Thanks and please keep continuing to make this kind of video!
Thanks for watching! I think you can go a couple different routes. You can work as an analyst or engineer and transition into one of the data science roles or you can pursue further education. If you were to work in one of the other roles first, you are making money and gaining work experience. I look at this as equally valuable to a masters degree. If you go the masters route, you are paying more money, but you will be able to check that box. If you have already worked as a data scientist at one company, the masters qualification is essentially waived for any other company you would go and work for. My suggestion is that if you can find a job you like as an analyst or software engineer, you should take it. If you don't find one that you are passionate about, it could be worth pursuing a masters degree. I hope this helps!
@@jason1224 Hi Jason, I am a recent data science undergraduate who landed a job in data science. I was in the same boat as you. I was worried if I'd be able to land in the data science field as an undergrad, but it's definitely possible. Of course, getting a master's degree will definitely be helpful as a lot of companies require at least a master's degree when selecting a candidate. However, it can be very expensive. You could potentially get into the field as a data analyst or a software engineer and work your way up. I've also made a video on this topic, so feel free to check it out!
You probably won't need to learn data collection, but you will need to have clean data. I would recommend learning eda as well to know what to put into the models.
In most videos I’ve watched, people say that you can’t start as a Data Analyst and later on become a Data Scientist. Plus, it was often mentioned that in order to be a Data Scientist, you need to have a Bachelors degree in the field. What do you think about that? I’d love to know! Thank you :)
I would generally disagree with them! I also disagree with the bachelors in the field statement. I would say only about half of the data scientists I've met or worked with had degrees in computer science or statistics and even fewer had a degree in data science (less than 1% since they are still very new). What matters most is your portfolio. That can open a lot of doors. Still, relevant education is helpful, but definitely isn't the whole story
Dear sir, I'm from Electronics background and I want to be Data Scientist, so I'm going for Masters of Data science please suggest were to start from..so I can have great knowledge about Data science and be a good Data Scientist.
What's the difference between someone with a masters or PHD in Math who is a data scientist and someone that doesn't have a math background and is calling themselves a Data Scientist because of their experience in the field and some of the classes they have taken in data science?
Honestly, the difference can be very big or very small. Having a masters or PhD doesn't guarantee that you will be a good data scientist by any means. I have met a few data scientists who don't have college degrees that I would argue are better than quite a few with advanced degrees. On the other hand, a masters or a PhD is a good heuristic for assessing if someone has the technical skills. These are generally asked for because it is harder to evaluate if someone who is self taught has the necessary capabilities.
@@KenJee_ds Thanks for the response! I agree just because you have a Math PHD doesnt always translate to being a great data scientist as there are other factors involved.
@@KenJee_ds Great video and response Ken. I would add that it depends predominantly on if the role is research focussed or not. Having worked as an analyst (the inside) and then also alongside hiring managers from the recruitment perspective ("outside"), they seldom consider candidates without PhD backgrounds for research focussed areas of machine learning. Ive seen extremely bright and accomplished data scientists still not make the cut for research ML roles especially within industries like quantitative finance. Rationale lies with appreciating the process of research they go through on their programs as their often looking for unique creative approaches to add to their own research process also. Just wanted to share the alternate perspective :-) great content!!!
I think that is generally a good starting point. If it is more interesting to you, you can actually also start as a data engineer or software engineer and move into a data science role later. I hope this helps!
2:20 i,m newbie to data field,but i dont think data cleaning is in responsibility of data engineer of course it,s vary from company to company but it,s not a part of data engineer job Data engineer mostly focus on data warhousing,data pipeline,etl &... if i am wrong i dont mind if you tell me
It really depends on the specific company. Most of the companies I was at, someone with a data engineering title was doing a large portion of the data quality control though. Honestly, I've had limited experience, so I wouldn't be surprised if this was not normal.
I really appreciate your content Ken. I have a question and will appreciate an answer from anyone .My background is in Petroleum Engineering and want to switch towards Data Scientist there is an online boot camp of 3 months from brainstation( 15000 $cad) to help me become a data scientist. Do you think it's worth it? Once you are a Data Scientist is it easy to transit to data engineer?
I think that this really depends on your personal situation. In theory you could learn the skills by self study for significantly less. I recommend watching this video where I compare the benefits and costs of bootcamps, certificates, and masters: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-Q9FjwzKFPuM.html
@@KenJee_ds Hey Ken! I decided to brush up my python skills beforehand. Guess there should be quite some time before I jump to MySQL again. Do share the link, hope to support your channel