Preparing Yourself for a Job in Data Science, Part 1: bootcamp
From Build a Career in Data Science by Emily Robinson and Jacqueline Nolis
Want a job in Data Science?
This article discusses one popular way to get the skills that you’re going to need: attending a bootcamp.
Going through a bootcamp
A bootcamp is an 8–15 week intensive course put on by companies like Metis and Galvanize. During the bootcamp you spend 8+ hours every day learning data science skills, listening to industry speakers, and working on projects. At the end of the course, you’ll usually present a capstone project to a room full of people from companies looking to hire data scientists. Ideally, your presentation gets you an interview and then a job.
Bootcamps teach you an incredible amount of knowledge in a short time. This means they can be great for people who have most of the skills needed for data science, but need a bit more. For instance, consider someone who has been working as a neuroscientist and has done programming as part of her work. A data science bootcamp could teach her topics like logistic regressions and SQL databases. With her science background plus those basics, she should be ready to get a data science job. Sometimes the best part of a bootcamp isn’t the knowledge itself, but the confidence you get from the program that you can do the work.
What you learn
A good bootcamp has a syllabus which is highly optimized to teach you exactly what you need to know to get a data science job and little more. This goes beyond technical skills and includes opportunities to work on projects and network with people. Here is more detail on what you should expect the program to cover.
Bootcamps are a great supplement to an existing education. For example, if you’re someone who has worked as a software developer for years, a bootcamp can quickly fill in the details you need around the math and statistical techniques and how to think about data. By doing a bootcamp you’ll be able to get a data science job quickly, without spending two years in a program like a master’s degree. This might be attractive if you already have a master’s degree in a non-data science field. The skills you typically get are:
- Introductory statistics — this includes methods for making predictions with data such as linear and logistic regressions, as well as testing methods you could use on the job like t-tests. Due to the limited time range, you won’t get deep into why these methods work, but you’ll learn a lot on how to use them.
- Machine learning methods — you’ll cover machine learning algorithms like random forests and support vector machines, as well as how to use them by splitting data into training and testing groups or using cross validation. You may learn algorithms for more specific cases like natural language processing or search engines. If none of those words made sense to you, that means you could be a good fit for a bootcamp!
- Introductory programming in R or Python — you’ll learn the basics around how data is stored in data frames, and how to manipulate it by summarizing, filtering, and plotting data. You’ll learn how to do the statistical and machine learning methods within the chosen program. Although you may learn R or Python, you probably won’t learn both and you may have to learn the other after you finish the bootcamp if you need it for your first job.
- Real-world use cases — You’ll not only learn the algorithms but also where people use them. Cases like using a logistic regression to predict when a customer will stop subscribing to a product, or how to use a clustering algorithm to segment customers for a marketing campaign. This knowledge is extremely useful for getting a job, and questions regarding use cases often show up in interviews.
Bootcamps have a highly project-based curriculum. Instead of listening to lectures for eight hours a day, most of your time is spent working on projects which best help you understand data science and gets you started with your own data science portfolio. This is a huge advantage over academia because your skills will be aligned with what you need to succeed in industry, because this is often similar to project-based work.
In a project, you’ll first collect data. This could be through using a web API which a company has created to pull their data. It could also be through scraping websites to collect the information off them or using existing public data sets from places like government websites. You’ll then load them into R or Python and write scripts to manipulate the data and run machine learning models on it. Then you’ll use the results to create a presentation or report.
None of those steps in the project require a bootcamp. This being said, having a project which is part of a bootcamp means you’ll have instructors guiding you and helping you if things go wrong. It’s difficult to stay motivated if you’re working alone and it’s easy to get stuck if you don’t have a person to call for help. That makes bootcamps valuable.
Lots of people go on from bootcamps to successful careers at places like Google and Facebook. The bootcamps keep alumni networks which you can use to get your foot into the door at those companies. The bootcamp may bring in data science speakers to talk to you during the program. People from industry view your final presentations. These people can also serve as connections to help you get a job at one of their companies. Having points of entry to companies with data science positions can make all the difference when it comes to finding jobs, and this perk of bootcamps must be stressed.
In addition to meeting people during the course itself, you can also use tools like LinkedIn to contact alumni from your bootcamp. They may be able to help you find a job at their company, or at least point you in the direction of a company which would be a good fit.
For all of these connections, you’ll have to be proactive. That means taking actions like going up and talking to speakers after they present and taking the initiative to send messages on social networks with people you haven’t talked to before. This can be scary if you aren’t comfortable with social interaction with strangers, but it’s necessary to get the value out of the bootcamp.
One significant downside to the bootcamp compared to self-teaching is the cost: the tuition is generally $15,000–$20,000. Although you may be able to get scholarships to cover part of the tuition, you also have to consider the opportunity cost of not being able to work full-time (and likely even part-time) during the program. Moreover, you’ll likely be on the job market for several months after your bootcamp. You won’t be able to apply during the bootcamp because you’ll be too busy and won’t have learned the skills yet, and even a successful data science job application process can take multiple months from application to starting date. That can end up being an aggregate of 6–9 months of unemployment, in addition to the cost of the program. If you’re able to teach yourself data science in your free time, or learn on the job, then you can keep working and not pay tuition, saving tens of thousands of dollars.
Choosing a program
Depending on where you live, there are likely only a few options for bootcamps. If you want to do an in-person bootcamp, even if you live in a large city there are probably only a handful of programs. If you don’t live in a large city and want to do a bootcamp, you may have to temporarily move to one. That can add to the cost of the program and make it more of an upheaval. Alternatively, there are online bootcamps for data science. Be careful, though: like with graduate programs, one of the benefits of in-person boot camps is that you’ll have people around you to motivate you and keep you focused. If you do an online course you lose that benefit, which can make an online bootcamp a $20,000 version of the same courses you could get through free or cheap massive open online courses.
In selecting between the bootcamps in your area, consider checking out their classroom, talking to some of the instructors, and seeing where you feel the most comfortable. But beware: with both academic degrees and bootcamps, there are lots of people looking to make a quick buck on people wanting to become data scientists. If you aren’t careful, you can end up completing a program which doesn’t help you get a job at all and leaves you with tens of thousands of dollars of debt. For bootcamps, it’s extremely important that you talk to alumni. Do you see successful graduates on LinkedIn? If so, talk to them and see how they feel about their experience. If you can’t find people on LinkedIn from the program it’s a huge red flag.
Data science bootcamp summary
Bootcamps can be great programs for people wanting to switch careers and already have some of the basics of data science. They can also be useful for people leaving school who want a few data science projects in their portfolio when on the job market. These aren’t designed to take you from “0 to 60” though; most of them have competitive admissions and you need to have a background in the fundamentals of programming and statistics to get in and then get the most out of it.
Stay tuned for part 2, in which we will discuss another important step in the quest to find a good data science job: building a portfolio.
That’s all for this article. If you want to learn more, check out the book on our browser-based liveBook reader here.