First, decide which data job you are aiming for
Data is not one job. A data scientist role usually asks for more mathematics, statistics and programming than an entry level reporting or analytics role. If you are changing careers, data analyst, business intelligence analyst or operations analyst is often the easier place to start.
For the rest of this guide, we use Data Analyst as the main example because it gives us a practical starting point. If you are aiming directly for data science, you will usually need deeper statistics and programming than the route below covers.
Before signing up for anything, spend an evening looking at real jobs. Search for one title, such as Data Analyst, and open 20 postings in your area. You will probably see the same tools and responsibilities appearing again and again. That becomes your first curriculum.
In our small sample, four of ten postings mentioned a bachelor’s degree or allowed experience as a substitute. None said computer science was the only acceptable field. That is encouraging, but it is a snapshot rather than a rule for the whole market.
| If you enjoy or already know | Consider | Why you might like it |
|---|---|---|
| Business questions, reports and spreadsheets | Data Analyst | A broad entry point for turning business questions into useful analysis |
| Dashboards, metrics and recurring reporting | BI Analyst | A good fit when you enjoy making information easier to monitor |
| Processes, forecasting and optimisation | Operations Analyst | Useful when you already understand how a business or service runs |
| Coding, statistics and modelling | Junior Data Scientist | Usually the more technical starting point and often the hardest first jump |
If none of these descriptions feels obvious, start by exploring Data Analyst vacancies in an industry you already understand. Your existing domain knowledge can help you ask better questions, even while your technical skills are still developing.
What we found in 10 current Data Analyst job postings
Before writing this guide, we looked at 10 current US job postings with Data Analyst in the title. We left out Senior, Staff, Lead and Director roles because they are not particularly useful when you are trying to break in. The sample includes general, people, product, marketing and operations analytics roles.
| Signal | Mentioned in | Context |
|---|---|---|
| SQL | 10 of 10 | Mentioned in every posting |
| Dashboards or BI tools | 9 of 10 | Tableau, Looker, Power BI or another BI tool |
| Prior experience or stated alternative | 9 of 10 | A course alone does not answer this screen |
| Python or R | 5 of 10 | Required or preferred in half of the sample |
| Statistics or experimentation | 5 of 10 | Required or preferred for selected roles |
| Excel or spreadsheets | 2 of 10 | Explicitly named, although use may be assumed elsewhere |
The clearest pattern was SQL: every posting mentioned it. Dashboard tools such as Power BI, Tableau or Looker came next. Python appeared too, but much less consistently. Only two postings explicitly named Excel or spreadsheets. Some employers may assume spreadsheet familiarity rather than list it, so we would not treat that small count as proof that spreadsheets are unimportant. Nine postings also asked for previous experience or an alternative, which is why you should build work you can discuss while you study.
Four postings mentioned a bachelor’s degree or allowed experience as a substitute. None said computer science was the only acceptable field. That is encouraging, but remember that this is a 10 job snapshot rather than the whole US market.
View the 10 postings in the sample
- SAMPLE
- 10 current US postings with Data Analyst in the title
- EXCLUDED
- Senior, Staff, Lead and Director roles
- CHECKED
- September 5, 2026
The skills worth learning first
Spend an hour searching for how to become a data analyst and you will end up with a ridiculous list of things to learn. You do not need most of them on day one.
Start with SQL because it appeared in every posting we checked. Learn enough to join tables, group data and test whether a result makes sense. Then choose one dashboard tool and get comfortable with spreadsheets. Add Python when it appears often in the jobs you want, or when your analysis becomes too repetitive to manage by hand.
Knowing the tools is not the same as having a portfolio. An interviewer needs to see what you did with them and why.
- One SQL project where you clean messy data, answer three business questions and explain your assumptions
- A dashboard built to answer one clear business question
- A short recommendation that says what the data cannot tell you
- A public case study another person can follow without your help
Do you actually need a certificate?
A certificate can be useful if you need structure. It gives you a syllabus, deadlines and a reason to keep going. What it cannot prove is that you can take a messy question, work through the data and explain what you found.
The IBM Data Science Professional Certificate covers Python, SQL and project work. That can suit someone who wants a broad introduction. If you are aiming at analyst roles, a shorter SQL and dashboard route may get you to a useful project sooner. Before paying for several months, compare the syllabus with the job list you made.
What should your first portfolio project look like?
Choose a question you can explain in one sentence. Use a public dataset with a clear licence, keep the cleaning steps and show the checks you made when something looked wrong.
Imagine a retail sales director wants to understand why margin is falling. The project does not need to look advanced. It needs to let another person follow your reasoning and understand what might change your recommendation.
- Question
- Which products and regions are hurting margin?
- Tools
- SQL and Power BI or Tableau
- Output
- Dashboard, queries and a one page recommendation
- What it demonstrates
- Data cleaning, business reasoning and clear communication
Keep a short readme, the SQL queries, notes about the data, dashboard screenshots and your conclusion in one project folder. A messy, honestly documented project is usually more convincing than a polished classroom exercise where the answer is already known.
One way to structure your first 16 weeks
This plan assumes roughly 8 to 10 focused hours each week. If you have only three or four hours, stretch the timeline instead of rushing through the projects. This is not a promise that everyone will be job ready in 16 weeks. Think of it as a structure for building the first pieces of work you can show and discuss.
After 16 weeks, the goal is not four certificates. It is a small set of work you can explain and defend. The questions and rejections you receive will tell you what to improve next.
When should you start applying?
You do not need to wait until you feel completely ready. Once you can confidently explain a couple of projects and meet a reasonable share of the requirements in the jobs you are targeting, start applying. For many analyst roles, that could mean using SQL to answer a question, building a clear dashboard and explaining the choices you made.
Treat the first applications as research as well as opportunity. Record which requirements appear repeatedly, where recruiters stop responding and which interview questions expose a real gap. If several interviews reveal the same weakness, that is more useful than guessing which skill to learn next.
What if the job asks for a degree or experience you do not have?
Read the wording carefully. Required and preferred are not the same, but neither should be ignored. If you meet most of the work requirements and can show relevant projects or experience from another field, an application may still be reasonable. Do not claim qualifications you do not have.
If nearly every suitable role asks for experience you lack, another certificate may not solve the problem. Look for a bridge: reporting work in your current team, a volunteer project with a real stakeholder, an internship, freelance analysis or an adjacent operations role. An accountant can bring financial context. A marketer can bring campaign knowledge. The portfolio should connect that experience to the analyst work you want.
Four of the ten postings we checked mentioned a bachelor’s degree or allowed experience as a substitute. None said computer science was the only acceptable field. That is useful context, but it is still a small sample and some employers will maintain firm degree screens.
Four mistakes worth avoiding at the beginning
Beginners lose a lot of time by trying to look advanced. These four detours rarely make the first application stronger.
- Learning machine learning too early Most junior analyst jobs are not waiting for you to build a neural network. Get comfortable with SQL, spreadsheets and dashboards first.
- Collecting certificates One good course can give you structure. Five certificates without independent work will not make your application much stronger.
- Building the same classroom portfolio project as everyone else Pick a less familiar dataset and answer a real question. The project should show how you think.
- Waiting until you feel ready to read job descriptions Start now. Current postings are one of the best free curricula available.
Is $120,230 a realistic starting salary?
No. The number at the top of this page belongs to the BLS Data Scientists occupation. It is a national median for workers at all experience levels, not a typical starting salary for someone finishing a 16 week analyst plan.
Data scientist is only one job in the wider data field, and it is not usually the easiest first role. Analyst pay varies with location, industry and experience, and the occupation can be classified differently across salary sources. Use current local postings and state wage data when you estimate your own range. Do not treat a national median as the return promised by a course.
If you are starting today
Pick one role. Read 20 current job descriptions. Learn SQL well enough to answer real questions, then add one dashboard tool. Build a project that gives you something specific to discuss in an interview.
You can add Python, statistics and deeper technical skills as the jobs you are targeting begin to require them. You do not need to learn the whole field before taking the first step.
Next, compare the syllabus of any course you are considering with the requirements you recorded from real jobs.
Where our numbers came from
Figures and provider details can change. We preserve the source and period so readers can distinguish national occupation data from provider claims. The ten job posting links and sample method appear in the research section above.
Sources reviewed September 5, 2026GOVERNMENT DATAU.S. Bureau of Labor Statistics: Data Scientists ↗Pay, typical education and 2025 to 2035 outlookCOURSE REFERENCEDIBM Data Science Professional Certificate ↗Current provider-published structure, duration and prerequisites