Biologist. Wanting to transition to Data Science. How? How long will it take?

llort_vbo wrote:

I transitioned from academia to data science and I went to a 3 month bootcamp. It was really helpful and I got s job shortly after. You do not need a master's. You'll pick up the stats that you need as you move on.

There are a lot of people spouting off about how much math you need and they're not right. Deep learning has advanced to the point that you can get by with a basic understanding of derivatives. Linear algebra is pretty helpful.

I agree with others that Python and SQL are essential. R is pretty easy to pick up once you've mastered those. Linked In has some decent free content for Python and I like Coursera/EdX/Udemy.

Good luck.

Agree with this perspective and would add the option to get a 1-2 year degree in bioinformatics is another option. I’ve been working in data analytics for 20+ years and currently lead a consulting and analytics shop for one of the largest companies in the AI space in the world. The keys to a successful career in data science are: good math and technical skills, passion for and deep content knowledge of certain data, ability to understand business problems and use data science to solve those problems; ability to make complex analytics simple and understandable to others. If you have these skills in spades and a few yrs of experience you should be earning ~$100k. Our senior data scientists are easily $120k+. Don’t have to live in Silicon or NYC. Think Austin, Research Triangle, Denver as examples of places with jobs.

iowakidscanrun wrote:

The keys to a successful career in data science are: good math and technical skills, passion for and deep content knowledge of certain data, ability to understand business problems and use data science to solve those problems; ability to make complex analytics simple and understandable to others. If you have these skills in spades and a few yrs of experience you should be earning ~$100k.

Cool. Now how do you suggest the OP, who probably hasn't taken math since high school, and has little experience with data, go about getting these skills "in spades"?

Try this wrote:

If you sant to make more money do more interesting work and you have a biochem degree, you really should look into PA (physician assistant) school or med school. I’m a surgeon our PAs all make about $120,000 in an inexpensive part of the country. They are surgical PAs. You can choose you’re specialty. It’s only 2 years of school for you (maybe less with your current degree). So, more money and more interesting work. Hours are 40/week. Strictly. Those who work more make more and can earn 200k pretty straightforwardly.

Geesus, I'm an Engineer looking for something less stress and that pays more for less hours. Can I switch to PA? You probably get your health care fully covered too, right? I'm mid 40s, is it too late to switch?

Why's there always posts here about people wanting become a developer? It's not glamorous. The best careers are in medicine.

This is an emerging field, and several schools are offering master's degree programs that you could look into. You'll want to know what pre-requisites are required for math/statistics and computer science skills are required, and you'll want to find what internships and job placement the school can help you with. Your best move is to combine the new data science skills with the Biology and Chemistry skills you have. As another poster said, you may find that you need to move to get into your first job in this field.

llort_vbo wrote:

I transitioned from academia to data science and I went to a 3 month bootcamp. It was really helpful and I got s job shortly after. You do not need a master's. You'll pick up the stats that you need as you move on.

There are a lot of people spouting off about how much math you need and they're not right. Deep learning has advanced to the point that you can get by with a basic understanding of derivatives. Linear algebra is pretty helpful.

I agree with others that Python and SQL are essential. R is pretty easy to pick up once you've mastered those. Linked In has some decent free content for Python and I like Coursera/EdX/Udemy.

Good luck.

I'm a mid-career research scientist and my friend at work (M.S. biology/microbiology) who's managed the lab across the hall for 10 years recently did a 3 month bootcamp. She's been applying for jobs in the Bay Area (23 and Me, Google, etc.) and going through intense all-day interviews that sound pretty grueling. She has yet to land a job, and she's got a lot of bioinformatics experience too. So it's not as easy as some folks are saying. Could just be ultra-competitive in the Bay Area though....

Actionable advice wrote:

iowakidscanrun wrote:

The keys to a successful career in data science are: good math and technical skills, passion for and deep content knowledge of certain data, ability to understand business problems and use data science to solve those problems; ability to make complex analytics simple and understandable to others. If you have these skills in spades and a few yrs of experience you should be earning ~$100k.

Cool. Now how do you suggest the OP, who probably hasn't taken math since high school, and has little experience with data, go about getting these skills "in spades"?

Well either you have the talent or you don’t. Either you seek out experience to broaden your skill set or you don’t. Nobody hands you a job for 100k. Get a bioinformatics degree, get a certificate, work with data in your current job, learn R and python,lots of entry points. The better your skills the more options you have and the more earning potential. I was listing the set of skills that determine level of success in this field. I don’t think that is a hard concept to grasp, is it?

Thank you to you both and some others who have given thoughtful, pragmatic advice. That's exactly what I was looking for, and it's very helpful.

Right now I'm starting a MOOC program in Data Science that's about 190 hours long. It gives an introduction to data science, python, statistics, SQL, etc. I figure that will give me a good start and idea as to whether I want to continue with the field or not. If I do, I can continue to put in another 500 hours or so, hone my skills a bit, and decide on a path forward as I get more acquainted with the field and what I like to do and/or what I am "good" at. I plan to focus pretty hard on statistics also and take a linear algebra course.

I know I'd have to move, but that's not a bad thing at all for me. I WANT to move. I don't really want to live in the Bay area, but there are plenty of other metro areas I would enjoy living. The jobs tend to be concentrated along the coasts, which is where I want to be anyway.

I know this will sound arrogant to a lot of people, but I'm very confident in my ability to hang and keep up with just about anyone. I am very aware I'm behind the curve as far as field-specific knowledge, but I know I can make up the difference if it's something I do want to do. I just don't personally know anyone in the field, nor am I the typical person trying to enter the field, so I wanted to come here for input and suggestions. Again, I really appreciate the responses.

Get a graduate degree; MOOCs are nice but I think in your situation you need more formal qualifications. Grad school allows you to make lots of connections, which will be key. Also, you will be more aware of the range of problems and research questions that data scientists address. It sounds like you are kind of fixating on data science as a "method" or "job" rather than as field. I'm not a data scientist so take what I'm saying with a grain of salt. I do work with and interact with data scientists on a daily basis, though. I guess the question you should think about is what are you going to do once you know all about data science? Which problems interest you and in what fields? Medicine, technology, biology, etc...Otherwise you can end up as a low-priced drone for the people who can design projects.

I didn't start grad school (MA and the PhD) until I was 32 and I am now a tenure-track professor at an R1 institution. Don't think you are too old for anything.

Upgrade wrote:

David S wrote:

Yeah I mean the OP would have a plausible story for wanting to get in to bioinformatics or computational genetics. I'm doing statistical genetics right now and it would be tough to get into it without a stronger background than the OP's, but plenty of people do come into it from other fields (I did, but my quantitative background was very strong). There are also data science positions at a lot of drug companies and so on.

I have thought about Bioinformatics... Several years ago I worked in research in a lab at a university that did genetic stuff and really enjoyed it. As someone with only a BS I didn't get that deep into it though. What is your job like? How do you like it? I don't really want to work for a drug company...

I like my job, but it's not exactly bioinformatics. As far as I can tell, most bioinformatics jobs involve developing or maintaining pipelines, making sure the data gets analyzed in the right sort of way, using the latest techniques to analyze data, and with a bit of data science and analysis thrown in. I think that can be interesting and fun, depending on the job.

My job is statistical genetics, which is more about trying to find ways to improve statistical models for genetics tasks. So basically you need to be able to understand enough about statistics and the state of the art in genetics to make improvements to how things are done. That can be relatively simple in cases where you're just trying to analyze a new type of data (like patient records or some new sequencing technique), but it can require a strong background if you're trying to improve on existing methods.

I wouldn't mind working for a drug company. You're helping to make people healthy. Yeah it's unfortunate that it's for profit, but that's how our system works. I dunno.

OP: Just out of curiosity, what's the last math class you took, and when?

Back to the basics wrote:

OP: Just out of curiosity, what's the last math class you took, and when?

I took Calc III in 2012. That's my last official math class. I fully admit I would need some refreshing of even quite low math subjects, though it wouldn't take much time at all to refresh my memory, at least with stuff up to Calc I.

Thank you, and I agree. I need grad school at the very least just to meet people and gain more experience. Work is good experience too, but it tends to be more narrow and repetitive than the experience people get in grad school (from what I can tell observing my friends). That's actually part of what has been frustrating to me the past several years. I know I am capable of doing so, so much more and WANT to do more, but that's just not really possible in my current position, nor is there any way for me to really move up in my current position for reasons I won't get into beyond saying having only a BS degree in science is very limiting almost no matter how skilled you are. However, I still want to do the MOOCs because I've found for me, they are MUCH more efficient because of how much faster I learn on my own than when in a program of some kind. I was actually in a masters program several years ago and left because of how slow paced and inefficient it was. There's tons of busy work and discussion classes where you just sit and listen to your peers share their (usually unintelligent) opinions, and you don't learn any actual skills. Not all programs are like that, but many are.

My plan is to continue with the MOOCs and then do a masters to get the formal credential. I figure having all the knowledge and skills from the MOOCs will make me somewhat of a powerhouse grad student and will allow me to really shine in grad school. I also have more time now to invest in learning these skills than I will have in grad school, so it makes sense, not to mention I'm making around $54,000/year now compared to the $20,000/year or less I will make in grad school.

Harambe's ghost wrote:

What's yer 5k pr?

14:13, though I beat guys in cross who went on to run 13:50s. I was injured in track both years I had stellar xc seasons.

I'm a mid-career research scientist and my friend at work (M.S. biology/microbiology) who's managed the lab across the hall for 10 years recently did a 3 month bootcamp. She's been applying for jobs in the Bay Area (23 and Me, Google, etc.) and going through intense all-day interviews that sound pretty grueling. She has yet to land a job, and she's got a lot of bioinformatics experience too. So it's not as easy as some folks are saying. Could just be ultra-competitive in the Bay Area though....

Sure, there's some issues regarding locale and I didn't mean to imply that it was easy for me to make this switch, but I think it's certainly possible for the OP, without an additional degree program. I saw other people with his/her education level come out of the program and land jobs. It takes a lot of work to get the right job and I would estimate that it generally took people 3-6 months of interviewing.

I didn't want to turn my post into a novel, so omitted details. After attending the bootcamp, I worked as an intern over the summer and went on 8 interviews. It took 3 months to get a job - part of that was learning to present myself better in interviews and part of it was I was looking for a senior position given my past experience.

I can't speak to your friend's experience as they may reflect the intense competition in the field. Some interviews were pretty rigorous.

hiding in the office wrote:

Get a graduate degree; MOOCs are nice but I think in your situation you need more formal qualifications..

I really disagree with this as it applies to data science. Where I hire, we are far less concerned with formal training than what's in your GitHub repo. I would strongly advise you to complete your MOOC and check out some Kaggle competitions. Have a couple of things in your repo that you can talk about during an interview. Show that you can clean and wrangle data, get some experience with cloud platforms, and that you can complete a project end to end.

uknow wrote:

You are a biologist, therefore

You are unable to understand statistical concepts.

You are poor at math, hence being terrified of logarithms and partial differential equations are something to do with cars.

You mention being good at excel... jesus wept.

You'll be no more than a glorified data clerk. Find some way to get at more "quantitative" things in your current field. Changing fields as suggested is unlikely to bring joy.

Stats is not that hard, get over yourself. I don't think much in mathematics is outside the ability of a reasonably smart person who tries.