I Wish I Had Known THIS Before Starting in Data Science

Key Takeaways

hello everyone ken here back with some more data science content for you today I'm talking about the things that I wish I had known before I started down the data science career path I'm going to talk a bit about the expectations related to the job as well as some of the tools that would have been useful for me to really get going full steam ahead as usual please hit that like button if this content is interesting to you and please subscribe if you'd like to see more videos like this now the first thing I'd like to mention is related to data wrangling data manipulation that's a lot larger part of the job then I expected going in most data scientists probably spend more than 50% of their time manipulating data and you know less than 50% of their time actually running algorithms it's doing data exploration etc so that's something that if you're really good at it can pay huge dividends in the long run and if you're quick and efficient at it that can also really add to your value and add to your efficiency another thing that was a little bit different than my expectation going in every day to science job that I've had is I didn't have as much control over the projects I was working on then I was expecting when you come out of school you're usually your assignments focus on projects that you choose and with a lot of the data science work really work at any company it's more of a push you're supposed to work on this because this aligns with business objectives so that's just something to keep in mind that is very common pretty much across to every type of job but if you want to be a go-getter you can also push for projects that are interesting to you I think it shows really good initiative to leadership that you want to work on specific initiatives and if you can have success there you might have more freedom going forward in your project selection etc now coming out of school I really thought I knew everything I was kind of a smartass and I started working and I learned that basically everyone that I worked with was really intelligent they knew what they were doing they knew a lot about data science and they knew about stuff other than data and I think one of the greatest things about working in this field is that you get to meet a ton of really intelligent people and you get to learn from them and that's a great asset to have there are people that come from physics backgrounds people that come from very technical computer science backgrounds and taking their knowledge and building it into your skillset is how you grow in this career and how you grow as a person now on the other hand you do work with people that aren't quite as technical there's always business stakeholders that don't have a good grasp of data science and it's an important skill to understand how to communicate with these people how to be able to take technical requirements how to take what's going on inside of an algorithm and explain them to to basically anyone now it's not that business people are less intelligent it's just that they have less focus and they haven't been studying this for a long period of time with learning from other people comes the feeling that data science is a lot bigger than you thought when I first started I kind of thought the data science was going well I have some data I run it through an algorithm I productionize it and I'm done data science can be whatever you want it to be it can be involving simulation you can bring in other computer science topics you can bring in a bunch of different types of modeling and it's constantly growing and changing the techniques that are used on a day to day basis and can even fluctuate and there's a wealth of research and things that are going on continuously that are improving the field so it's important to continually keep growing to keep learning to see what's on the cutting edge there's also new technologies that are constantly released that can make your life easier and can also change the way that you think about different problems now in terms of tools I think there are a couple of things I really wish I had put more upfront effort in in learning there's the traditional stuff like Python a bunch of different sequel things or dealing with databases relational etc but there's another element of production ization of code that is generally overlooked people use things like docker they use sage maker and these things are being more and more commonly used in the field and it's important to have a grasp of them you can write really good code but if it's not generalizable if it's not your if you're not able to make it an endpoint it's not super useful in the field so I would focus on learning different frameworks that are out there for production iliza code and production Eliza's in data science while tools are really important at the end of the day work revolves around other people and in every company the culture the way data science is done is very different so the way that projects are scoped the way that projects are run the way that requirements are gathered really very significantly by company and you can work in they vary by the size of the company if it's small or large what stage the company is at if it is a startup or a high-growth company or an established corporation so learning about how the companies that you're applying for positions at how they operate is extremely important and you have to find a culture that really that really meets your style some people work on a bunch of different projects at the same time some people have a very linear focus in their in their projects and the things that they do on a day to day basis so make sure you do your homework make sure you really research the type of organization that you are applying to and getting into and to that point to a certain extent office politics do play a fairly large role in data science data science sits in between engineering and software and the business side and so there is a lot more communication and kind of jockeying for time than there are in some other positional roles so I think that that's just something that's an inevitability but you have to be prepared for it you're going to be asked to do a bunch of different things you're going to be pulled different ways by different people and it's really important to make sure that your manager the people that you're working for have your best interest in mind I think that you find this out relatively quickly through the interview process and that's one of the questions that I generally ask when I'm interviewing I ask what the structure is like who are our main stakeholders and how does my perceived boss react with those stakeholders and you know how does my work that shown etc so that's something to keep in mind that's something to focus on because in some companies office politics can be really bad and some they exist but they are beneficial for you in your role one of the other biggest deviations from education - in practice data science is the idea of statistical significance in business you're just trying to create positive expected value rather than to have certainty that something works you want to evaluate what the return is compared to what the cost is if the cost is that things stay the exact same if your algorithm doesn't have any impact that's generally fine if you want to do things that like can possibly create value now I'm not saying that statistical significance isn't important in some areas you know if you're testing a new drug if you're in the pharmaceuticals industry you really want to be sure that people are getting sick but if you're focusing on different customer groups to target if you can get a 1% better outcome that might be worth putting into production now people really get wrapped up in this but we forget the statistical significance nowadays with the size of data that we have is relatively arbitrary we just want to be able to create a positive net effect and that really goes a long way with the business some people get too theoretical but at the end of the day the sometimes the the best we can do is to create a 1% lift and that can create millions even billions of dollars of revenue which is pretty cool thank you so much if you have any questions relating to any of the topics I talked but please leave them in the comments section as usual have a great one and thank you for watching