A hiring manager once told me she had close to 300 applications for one data analyst opening, and most of the applicants were quietly applying to be data scientists without realising it. They had spent months on machine learning tutorials and almost no time on the spreadsheet and SQL work the job actually asked for.

That gap, between the title people chase and the work the job needs, is exactly where this series begins. If you are aiming for your first data job, the most useful thing you can do in week one is get clear on which of three very different roles you are actually training for. This part draws that map.

What a data analyst actually does

Strip away the job-ad language and a data analyst does one thing: turns raw data into an answer a non-technical person can use. Someone in the business has a question, sales dropped last month and nobody knows why, or which product line actually makes money, and the analyst goes and finds out. That means pulling the relevant numbers, cleaning them up so they are trustworthy, looking for the pattern, and then explaining the pattern in plain words with a chart or a dashboard behind it. A dashboard, if the word is new, is just a live screen of the key numbers that updates on its own so people can check it without asking you.

Most of the day is unglamorous, and that is worth saying up front. Analysts spend a large share of their time cleaning data, which means fixing dates stored three different ways, removing duplicate rows, and reconciling two systems that disagree about the same customer. The famous line in this field is that data work is roughly 80 percent preparation and 20 percent analysis, and while the exact split varies, the spirit is right. The insight at the end is the reward, but the preparation is the craft.

The tools reflect that work. Analysts live in spreadsheets such as Excel or Google Sheets, they query databases with SQL, a language for asking a database questions that we cover from scratch later in this series, and they build reports in a business intelligence tool such as Power BI or Looker Studio. Business intelligence, usually shortened to BI, simply means software for turning company data into charts and dashboards. Notice what is not on that list: you can do the core of this job without writing a single machine learning model.

Analyst, scientist, engineer: who owns what

The cleanest way to keep the three roles apart is by the question each one answers. The analyst answers what happened and why. The scientist answers what is likely to happen next. The engineer answers a more basic question that comes before both: can we even get at this data reliably. Hold those three questions in your head and almost every confusing job title sorts itself out.

Here is the same split laid out with the tools and the typical output for each. Read it as a starting sketch, not a rigid rulebook, because real teams blur these lines constantly. A machine learning model, for the table below, is software that learns patterns from past data to make a prediction, and a data pipeline is the automated plumbing that moves data from where it is created to where it can be analysed.

A starting sketch of the three roles. On small teams one person often wears two of these hats at once.

flowchart TD
  Q[A business question arrives] --> C{Is clean data available}
  C -->|No, not yet| DE[Data engineer builds the pipeline]
  DE --> C
  C -->|Yes| K{What kind of answer is needed}
  K -->|What happened and why| DA[Data analyst]
  K -->|What is likely next| DS[Data scientist]

Where the three roles overlap

The tidy table hides a messier truth: these roles share a lot of ground, and the boundaries move from company to company. Cleaning data, writing SQL, and building charts show up in all three jobs. A data scientist still cleans data before modelling. A data engineer still writes queries to check a pipeline is correct. So the skills you build as an analyst are not a dead end, they are the shared foundation the other two roles stand on. That is a big part of why starting as an analyst is such a sensible on-ramp.

Titles make this worse before they make it better. A business intelligence analyst, a reporting analyst, a product analyst, and a marketing analyst are, at the core, all doing analyst work with a different subject in front of them. A business analyst, confusingly, is often not a data role at all and leans more toward requirements and process. When you read a job posting, ignore the noun in the title for a moment and look at the verbs in the responsibilities. The verbs tell you the truth about the job.

There is also a natural career path hiding in this overlap. Plenty of data scientists and analytics engineers started as analysts, learned the statistics and the programming on the job, and grew into the next role once they had the foundation. You are not locking a door by starting as an analyst. You are opening the one that most reliably leads to the others.

What the roles pay and how fast they are growing

Beginners always ask about money, and they should, so let me give you real numbers rather than a vague reassurance. The United States Bureau of Labor Statistics, the government body that tracks jobs and wages, publishes median pay and growth projections for data occupations. The median wage is the point where half of workers earn more and half earn less, which is a fairer middle than an average that a few huge salaries can pull upward.

The table pulls the closest official occupations to analyst and scientist work. For scale, the median wage across all United States jobs was 49,500 dollars in May 2024, and the average growth rate across all occupations for 2024 to 2034 is 3 percent. Every role below clears both bars comfortably.

Figures from the United States Bureau of Labor Statistics Occupational Outlook Handbook, May 2024 wages and 2024 to 2034 projections. Data analyst work is spread across several of these codes, which the next callout explains.

Why analyst is the right first role

If your goal is a first data job, aim at the analyst role first, and I say that for three concrete reasons. First, the barrier to entry is lower in the honest sense: the core skills are spreadsheets, SQL, and a BI tool, all of which you can learn without a maths or computer science degree. Data science, by contrast, usually expects real statistics and programming before anyone will hire you, and data engineering expects software and cloud skills. Second, there are simply more analyst jobs, because every department that has a spreadsheet has questions, while predictive modelling is concentrated in fewer teams.

Third, and this is the part beginners miss, the analyst skill set is the base camp for everything else in data. The SQL you learn to answer what happened is the same SQL a scientist uses to pull training data and an engineer uses to test a pipeline. Start as an analyst and you are not choosing a lesser path, you are choosing the path with the widest set of doors at the end of it. You can always add statistics and Python later, and this series will start you on both when the time is right.

Two myths that stall beginners

The first myth is that you need advanced maths to become a data analyst. You do not. The maths an analyst uses day to day is averages, percentages, ratios, and a handful of ideas we cover gently later in this series, such as what a median is and why an outlier can mislead you. If you can work out a discount and read a bar chart, you already have the numeric base. The deeper statistics belong to the data scientist, and you can pick them up later if you want to move that way.

The second myth is that you must learn Python before anyone will hire you, or that AI tools have made the analyst job disappear. Neither is true. Plenty of working analysts do most of their job in spreadsheets, SQL, and a dashboard tool, and add Python only when it saves them time. As for AI, it writes queries and drafts charts faster than before, but someone still has to ask the right question, judge whether the answer is trustworthy, and explain it to a room of people. That judgement is the analyst, and it is the part of the job that is getting more valuable, not less.

Aim for analyst, and start with the questions

Here is the recommendation for this part, plainly. If you are starting out, train for the data analyst role, not the data scientist role, and do not worry yet that the analyst title pays a little less on paper. The analyst job is more reachable, there are more of them, and it is the foundation the higher-paying roles are built on. You can grow into a scientist or an engineer once you have the base, and many people do exactly that.

The one habit to build starting today costs nothing: when you meet any number, sales, scores, prices, ask what happened and why before you ask anything else. That single reflex is the whole job in miniature, and everything else in this series, the spreadsheets, the SQL, the dashboards, is just learning to answer it faster and more convincingly. In the next part we map the full analyst toolkit and lay out a skills roadmap, so you know exactly what to learn and in what order.

Your task this week: pick one number from your own life or work, a monthly bill, a step count, a sales figure, and write down two honest questions about what happened and why. That is your first analysis, and you already have everything you need to start it. For the full path from here, keep the Data Analyst guide open as your map.

References

• Data Scientists, Occupational Outlook Handbook, U.S. Bureau of Labor Statistics
• Operations Research Analysts, Occupational Outlook Handbook, U.S. Bureau of Labor Statistics
• Market Research Analysts, Occupational Outlook Handbook, U.S. Bureau of Labor Statistics

About The Author

Dr. Pranay Jha

Dr. Pranay Jha is a Cloud and AI Consultant with 18+ years of experience in hybrid cloud, virtualization, and enterprise infrastructure transformation. He specializes in VMware technologies, multi-cloud strategy, and Generative AI solutions. He holds a PhD in Computer Applications with research focused on Cloud and AI, has published multiple research papers, and has been a VMware vExpert since 2016 and a VMUG Community Leader.

See author's posts