The world’s leading publication for data science, AI, and ML professionals.

Skills and Projects You Need to Get a Data/Business Analyst Position 2021

Part 2 of a New Graduate's Guide to Becoming a Data/Business Analyst

https://elianag.medium.com/skills-and-projects-you-need-to-get-a-data-business-analyst-position-2021-d5c21c88e3bd
https://elianag.medium.com/skills-and-projects-you-need-to-get-a-data-business-analyst-position-2021-d5c21c88e3bd

Contents

Projects SQL Website

Projects

You’ve probably heard a lot about the importance of projects.

Projects are particularly crucial for a recent graduate applying to be a data analyst or a Business Analyst because they demonstrate proficiency in skills that you probably didn’t use outside of a classroom context.

As much as I might want to list my Data Structures class as experience because I used Java during it, please for the love of god, do not put any computer science labs or classes on your resume. Not only is putting labs on the internet likely directly against your school’s honor code, but it also looks amateur. The exception to this is that you can list classes under a Course Highlights section on your resume that is located beneath your degree.

What makes a good project?

  • Cool, uncommon dataset
  • Clear, answerable question
  • Visualization and statistical techniques used to answer the question
  • Some form of an answer to your question

There are some crazy talented people on the internet making insane and complicated data science projects. Keep in mind that most of those people have Master’s degrees, PhDs, and tons of work experience. Rest assured, you do not have to be one of those people to get a job.

You’ll notice that I didn’t mention any machine learning techniques. It’s because you do not have to know machine learning to get an entry-level Data Analyst job. If you’re lucky, they’ll let you play with machine learning while you’re on the job. The data scientists you work with will be doing most of the actual machine learning.

The analytical techniques you use don’t have to be anything crazy, at least for a data analyst or business analyst. In fact, you don’t even have to include any sort of machine learning techniques. Just make some pretty visualizations of the data that tell you something that isn’t immediately obvious when you look at a spreadsheet.

That said, if you want to use machine learning, go for it! Machine learning is fun and there are lots of ways to integrate it into your projects.

How do I find a dataset?

There are countless sources of datasets. Kaggle and data.gov are a couple of good ones. Most datasets are going to be downloadable as .csv files, which is an easy file format to work with in most languages and applications.

If you’re interested in machine learning, you should know about some of the classic datasets. You can find them at the UCI machine learning repository.

Keep in mind: Try not to use any of the UCI datasets for your projects. They are extremely commonly used, which means that people have analyzed them to death. They’re mainly useful as a means of comparing machine learning techniques in academic papers. But as a machine learning person, you should know about them.

My recommendation:

Instead of looking through existing datasets, I would think about what kind of data you want to work with. After you come up with an idea, specifically search for a dataset. You’ll find less commonly used datasets and end up with a more interesting project.

Ask yourself:

  • What’s something I’m curious about? What’s something I’m passionate about?
  • What’s something I know something a little bit about already that others don’t know anything about?
  • Do I want to work primarily with quantitative data or _qualitativ_e data?
  • What skills do I want to prove that I know or learn through this project?

If you have the programming skills, you should feel free to scrape your own data from the internet! It’s not that difficult and there are lots of tutorials that will tell you exactly how to do it.

Make sure the data is quality – before you touch it, look for the ratio of missing data to total data, and total lines of data. A little missing data is fine, but a lot of missing data is a problem. If you find a dataset that has less than 1000 rows, it’s probably not going to be able to tell you anything interesting or useful.

How do I come with a question?

You’re not going to finish your project or do it well if you don’t really care about the answer to your question. So pick a question that genuinely interests you!

Also – don’t feel bad if it takes a while to come up with a question! I would argue that counter-intuitively, coming up with a question is often the hardest part of the entire project.

This is why I would recommend against using Kaggle as a resource for a personal data project unless you’re going hardcore for Data Scientist jobs. Kaggle usually defines the question for you, which basically means that Kaggle does the hardest part for you.

Luckily, since you just completed an entire undergraduate degree, you probably have some experience coming up with good questions!

Think about the last time you came up with a project in school without much guidance from professors or wrote a thesis statement for an essay based on papers you read in class. The skills that helped you craft good arguments and completable projects are the same skills that you’re going to use right now to come up with a question.

Examples

The project can be anything from a random thought once you had to something you’ve spent years thinking about. One day, I was wondering what people complained about on the internet. I taught myself about some relevant natural language processing techniques and had a complete project within a week. You can see it here.

Another example: I knew I wanted to do a project that had something to do with politics since I was a politics major in college and I wanted to teach myself how to use Tableau. So I searched for a dataset, found some cool ones at openelections.net, and ended up with this project.

Final thoughts on projects

  • Don’t be afraid to do things that you’ve seen other people do. If it demonstrates your skills and you do it in a different way than other people, it’s okay to come to the same conclusion as other people. Your future employer cares a lot more about how you think and work through a problem than a unique conclusion.
  • Try to have at least 3 different projects that demonstrate your skills, preferably different skills in different projects.
  • If you did a Data Science-related project in school that was entirely your idea, you should feel free to list it in the projects section of your resume. For example, my course recommender project was something I originally did for an Artificial Intelligence class.

SQL – what’s the deal with SQL and how do I get good at it?

SQL is a skill that many, but not all, employers hiring for data analysts and business analysts are going to want you to know well and will likely test you on during the interview process.

SQL is a logic-programming language, which basically means that every query is a logical statement in pseudo-English that tells the computer exactly what subset of the data __ you want. In general, you’re not going to be iterating over anything. This is different from many programming languages like Python, Java, and C.

Anyone who took Discrete Mathematics in college should know more than enough set theory to understand and write SQL logic. If you didn’t take Discrete Mathematics or don’t feel comfortable with set theory, I’d recommend that you go through an online course on basic set theory before you even start writing SQL queries. It’ll make your life so much easier.

By the way – you shouldn’t bother including any SQL in your projects unless you really want to. If a potential employer really cares that you know SQL, they’ll give you a timed test. Honestly, I have a lot of test anxiety, so I didn’t exactly ace those tests. Given that, aspects of this section are my best guess at how good at SQL you need to be, rather than hands-on experience.

So, how do I learn and get good at SQL?

There’s a ton of versions of SQL out there that all have slightly different syntax and keywords. MySQL is the most common, so I’d recommend focusing on learning that.

There’s a lot of resources out there to learn and get good at SQL – CodeAcademy, W3Schools, and more. I don’t really have a preference for one over another but you probably will.

To practice, I found LeetCode to be really useful. LeetCode is mainly used by people trying to practice for Software Engineering interviews, but they have tons of SQL problems on there that were quite similar to ones I saw on actual tests. Drill "easy" and "medium" questions until you can consistently correctly answer them in 5–10 minutes. Throw in a few "hards" to challenge yourself. Be sure to time yourself – it may make you feel bad initially, but you will improve and it’ll give you a lot of insight into how proficient you are.

LeetCode is free for limited problems or $35 per month to get access to all of the problems. Frankly, I think the $35 is worth it if you are really serious about practicing.

If you don’t want to pay anything, there are many free resources to practice online. You also can download MySQL for free onto your computer if you want to practice a version of the real thing, especially if you want to practice creating and deleting tables and rows.

Here’s an example of real SQL interview questions that you can look over.

Specific keywords/functions to know:

  • JOINS (all of them, and how to use them quickly and proficiently)
  • WITHDRAW
  • GROUP BY
  • ORDER BY
  • the difference between WHERE and HAVING
  • PARTITION
  • the difference between UNION and UNION ALL

This article and this article have some good tips on this as well.

To be fair, I’m not really the best person to be telling you how to get good at SQL, but this is what I would be doing if I had found a job a month later than I actually did.

Website

Do you need a website to get a data analyst job? No. Does it help? Absolutely.

Advantages of a website:

  • It shows potential employers that you have an eye for visual design and that you can communicate well
  • Enables you to control how you appear on the internet
  • It’s easier to highlight and summarize projects on a website than on GitHub or Medium
  • Can host your personal blog, if you have one

You can think of your website as a nicer, better version of your LinkedIn page – I would recommend linking to it from your LinkedIn page. The major advantage of a personal website over a LinkedIn page is that you can customize it precisely to your liking.

For an example, feel free to check out my website. If you want to find others, just search "data science portfolios" or message me and I can send you a few links to some portfolio websites that I saved.

Your website will particularly impress business people. You will inevitably interview with business people if you make it far enough in the interview process, so impressing them matters. It also gives you an easy way for people to check out all of your work without you having to send them multiple links.

Be aware, however, that your hiring manager will probably not take it as any indication of your technical skills, even if the business people do.

What should I put on my website?

  • A landing page that contains information the first thing that you want any potential employer to know about you
  • Links to all of the websites you want to highlight
  • About Me section
  • Projects, including a summary and relevant links
  • Blog (optional) – blogs are a great way to talk about your projects, so I’d recommend one, and it’s perfectly ethical to cross-post on your personal website and Medium
  • Resume (optional)

How should I make it?

Just use WordPress. It’s okay, even if you have a CS degree, to make it in WordPress. It’s more important to have a nice-looking website than hard-coded HTML.

Use a custom domain, if you can. Ideally, it’s your first and last name followed by ".com" (i.e. mine is elianagrosof.com). If that’s not available, feel free to get creative about getting as close as you can to your actual name.

You’ll need to buy the domain (usually about $20 per year) and likely pay for some sort of hosting service, which should also be inexpensive.

Where should I link to my website?

  • On your LinkedIn page
  • On your resume (along with a link to your LinkedIn page and GitHub)
  • In the link spots for "Portfolio" and "Other" on some job application forms

— – – –

A New Graduate’s Guide to Becoming a Data/Business Analyst

Part 1: How to Become a Business Analyst 2021 Part 2: Skills and Projects You Need to Get a Data/Business Analyst Position Part 3: How to Network and Use LinkedIn to Find a Job Part 4: How to Use Professional Groups to Find a Business Analyst Job 2021 Part 5: How to Keep Your Sanity During Your Job Search


Related Articles