How to Break Into Data Without a Computer Science Degree

Nobody in my family worked in technology. I grew up in a household in Evansville, Indiana, that lived on $776 a month. I didn't know anyone building companies or working with data. Success felt like something that happened to other people.

So when someone tells me they want to get into data but "don't have the background," I take it personally, in the best way. Here's the honest roadmap.

First, the myth

The myth is that data jobs belong to people with computer science degrees from famous schools.

The reality is that a huge share of great analysts came from somewhere else: teaching, nursing, retail, logistics, the military, customer service, social work. And they're often better at the job, because data work is mostly about asking the right questions of a business you understand.

A degree can help. Proof helps more. In my experience mentoring people into data roles, a candidate who can walk me through a project they built, with messy data they cleaned themselves, beats a candidate with three certificates and nothing to show.

The skill stack, in order

Learn these in sequence. Each one makes the next one easier.

1. Spreadsheets, for real (2 to 4 weeks)

Not just typing numbers. Learn sorting and filtering, XLOOKUP (or VLOOKUP), SUMIFS and COUNTIFS, pivot tables, and basic charts. Most businesses still run on spreadsheets, and pivot tables teach you the core idea of all analytics: grouping and summarizing.

2. SQL (1 to 3 months)

SQL is the language for asking questions of databases. It's the single most valuable skill on this list, and it's very learnable. Focus on:

• SELECT, WHERE, ORDER BY, the basics of pulling data
• GROUP BY with COUNT, SUM, AVG, the same idea as pivot tables
• JOINs, combining tables. This is where most beginners get stuck, and where most real work lives. Practice until you can explain the difference between an inner join and a left join to a friend
• Window functions like ROW_NUMBER() and running totals, once you're comfortable

A tip from years of doing this: when a join gives you more rows than you expected, stop. You've almost certainly joined on a key that isn't unique. Learning to notice that is what separates analysts from people who just run queries.

3. One visualization tool (1 month)

Pick one, like Power BI, Tableau or Looker Studio, and get decent at it. Learn to build a simple dashboard with filters. More importantly, learn to choose the right chart and to put the most important number where the eye lands first.

4. A little Python (ongoing)

Once SQL feels comfortable, learn enough Python with the pandas library to load a CSV, clean it, reshape it and make a chart. You don't need to become a software engineer. You need to automate the cleanup you'd otherwise do by hand.

5. The skill nobody lists: data skepticism

Every dataset is lying to you a little. Missing values. Duplicates. A field called "date" that's actually three different date formats. A survey that only reached people with internet access. Great analysts ask "how was this collected, and who's missing?" before they ask anything else. I built a whole framework around this, called the Data House, because so many projects fail on the foundations. You can read it in Get Your Data House in Order Before You Buy AI.

Build three portfolio projects

Projects are your résumé. Make three, using real public data about something you actually care about. Caring shows.

Ideas that work well:

• Your city. Seattle publishes a ton of open data at data.seattle.gov. 311 requests, permits, traffic. Ask a real question: Which neighborhoods wait longest for pothole repairs?
• Your industry. If you worked in retail, analyze public retail sales data. If you were a teacher, use public education data. Your domain knowledge will shine.
• Your community. Census data on internet access by neighborhood. Small business data. Something that matters to people you know.

For each project, write a short page with four parts: the question, the data, what you found, and what you'd do next. Include one honest limitation. Hiring managers love seeing that you know what your analysis can't tell them.

Landing the first role

Look for "adjacent" titles. Your first data job might not say "data analyst." Look at operations analyst, reporting specialist, business analyst, program data coordinator, marketing analyst. Plenty of great careers start there.

Start where you are. The easiest data job to get is often at the place you already work. Volunteer to build the report everyone complains about. Automate a spreadsheet. Become the person people bring questions to.

Nonprofits and government need you. Mission-driven organizations are often desperate for data help. It's great experience and it matters.

Talk to people. Go to meetups. Ask analysts for 20-minute conversations about their work. Most of us remember being new and are happy to help.

Practice explaining. In interviews, you'll be asked to walk through a project. Practice telling it as a story: the problem, what you tried, what went wrong, what you learned.

What about AI?

You've probably heard AI will replace analysts. Here's what I see: AI is making the typing part faster. It writes a first draft of a SQL query in seconds. But someone still has to know whether the query is right, whether the data is trustworthy, and what the answer means for the business.

That's the job. AI makes analysts who understand the fundamentals more powerful, and exposes the ones who don't. Learn the fundamentals.

You belong here

If you're reading this from a kitchen table after a long shift, wondering if this is realistic for someone like you: it is. Start with one spreadsheet skill this week. Then the next one.

Pour me another: What's stopping you from starting? Tell me through the contact form. Your answer might become my next class.

Sip slow, stay curious.D