You can do most of a data analyst’s job with four tools, and not one of them is machine learning. That surprises people who show up expecting to write clever algorithms on day one. The real toolkit is a spreadsheet, a query language called SQL, a dashboard tool, and, later, a little Python.
Learn those four in the right order and you are hireable in a few months. Learn them in the wrong order, or chase the flashy one first, and you stall for a year. This part lays out the toolkit and the exact sequence I would follow if I were starting today.
Key takeaways
The core toolkit is four things: a spreadsheet, SQL, a business intelligence tool, and Python. You do not need all four to get your first job. The first three carry most analyst work.
Order beats speed. Spreadsheets first because they give instant feedback, then SQL because it is the skill most postings demand, then one dashboard tool, and only then Python.
A realistic part-time plan runs about four months. Everything in this list is free to download or use, so cost is not your blocker. Sequencing and practice are.
The four tools you actually need
The toolkit is smaller than the internet makes it look. Here are the four pieces in one plain sentence each. A spreadsheet, such as Microsoft Excel or Google Sheets, is a grid where you type data and write small formulas to calculate on it. SQL, said as either "sequel" or "S Q L", is a language for asking a database a question and getting rows back. A business intelligence tool, BI for short, is software that turns a table of numbers into charts and a live dashboard other people can read. Python is a general programming language that, with a few add-on libraries, automates the repetitive parts of all of the above.
Think of them as four layers stacked from friendliest to most technical. Each layer teaches an idea the next one reuses, so the effort compounds instead of resetting. The diagram shows the stack and what each layer is good at, so you can see why the order is not arbitrary.
Spreadsheets come first, and here is why
Beginners want to skip spreadsheets because they feel too ordinary to matter. That is a mistake. A spreadsheet gives you something no other tool does at the start: instant feedback. You type a number, write a formula, and see the result change in the same second. That tight loop is how you build a feel for data, and that feel is the thing every later tool depends on. Most business data still lands in your inbox as a spreadsheet anyway, so the skill pays off from week one.
Keep your spreadsheet study focused. Learn the common formulas, learn lookup functions such as XLOOKUP that pull a value from another table, and learn pivot tables, which are a drag-and-drop way to summarise thousands of rows into a small readable table without writing a formula. Add basic charts and you can already answer real questions. Do not spend six months mastering every macro. Once pivot tables feel natural, you have taken what spreadsheets have to teach and it is time to climb to the next layer.
SQL is the skill that gets you hired
If you learn one technical skill from this whole series, make it SQL. A company’s real data does not live in a spreadsheet, it lives in a database, which is an organised store of tables that can hold millions of rows. SQL is the language you use to ask that database a question, for example show me every order from last month over 500 dollars, and get the matching rows back. Nearly every analyst job posting lists SQL, and many interviews are built around a live query test. This is the skill that moves you from playing with sample files to working with the company’s actual numbers.
The good news is that the useful core of SQL is small. You need to select and filter rows, join two tables so you can combine related data, and group rows to compute totals and averages. That is most of the job. You do not need to learn how to design or administer a database, that is data engineer territory. We build SQL from an empty query up across Parts 6, 7, and 8 of this series, using the same free databases the professionals learn on, so you do not have to find your own starting point.
What a BI tool adds
A query answers a question once. A business intelligence tool answers it forever. You connect the tool to your data, build the charts once, and it refreshes on its own so a manager can check the dashboard on Monday without asking you to rerun anything. The two most common free starting points are Power BI, a free desktop application from Microsoft that you download, build reports in, and publish to a shared web service, and Looker Studio, a free web-based tool from Google, formerly called Google Data Studio, that runs in the browser and connects to sources such as Google Sheets and many others.
Pick one and go deep, do not sample all of them. The concepts transfer, so once you can build a clean dashboard in Power BI you can rebuild it in Looker Studio or Tableau in an afternoon. Choose Power BI if you are aiming at companies that run on Microsoft, and Looker Studio if your data already lives in Google products or you just want the fastest free start with nothing to install. We build your first dashboard together in Part 14.
Worked example
Your manager asks every Monday, which five products sold best last week. Without a BI tool you rerun a spreadsheet and email a screenshot each time. With one, you write the logic once: connect to the sales table, sort by units sold, keep the top five, and drop it on a dashboard. Next Monday nobody emails you at all. They open the link and the numbers are already current. That shift, from answering the same question weekly to answering it once, is the whole point of the layer.
Where Python fits, and when to wait
Python is the fourth layer, and it earns its place the day a task is too big or too repetitive for a spreadsheet and too fiddly for a single query. The tool most analysts reach for is pandas, a free Python library for working with tables in code, where a table is called a DataFrame. With pandas you can clean a messy file, join data, and produce a chart in a script you can rerun on next month’s data with one command. The library is mature and widely used, with version 3.0.4 current as of mid 2026, so what you learn now will keep working.
Here is the trap: do not start here. Python is the most common way beginners waste their first three months, because it feels like real programming and looks impressive. But an analyst who cannot write a join in SQL and cannot build a clean chart will not pass an interview no matter how much Python they know. Learn Python after you can already answer questions with spreadsheets, SQL, and a dashboard. At that point pandas amplifies skills you already have, which is exactly when it is worth the effort. We start pandas gently in Part 15.
None of this costs money to start
One worry I hear a lot is that real tools must be expensive. They are not. Every layer in the toolkit has a free option that professionals actually use, so you can start today without spending anything. Google Sheets runs free in a browser, and Excel has a free web version too. For SQL, both PostgreSQL and MySQL are free and open source, and they are the same engines large companies run in production, so you are not practising on a toy. Power BI Desktop is a free download from Microsoft, and Looker Studio is free to use in the browser with nothing to install. Python and pandas are free and open source as well.
So the only thing standing between you and this whole roadmap is time and a plan, not a budget. Do not let an imagined paywall stop you, and do not pay for a bootcamp before you have tried the free path and hit a real wall. Spend money later, once you know exactly which gap you are filling. Until then, treat the free tier as the job itself, not a stripped down preview of it.
A roadmap you can finish in a few months
Now the sequence, drawn as a path so you can see where each skill hands off to the next. Follow the arrows. Each stage assumes you finished the one before it, which is why skipping ahead hurts more than it helps.
The table puts rough timings on that path for someone studying part time, a few hours most evenings. These are targets to steer by, not a test. Move faster if a stage clicks, and give a hard stage the extra week it needs without guilt.
| Stage | Rough weeks | Main tool | What you can do after |
|---|---|---|---|
| Spreadsheets | 3 | Excel or Google Sheets | Summarise data with pivot tables and charts |
| SQL | 4 | PostgreSQL or MySQL | Pull and combine real data from a database |
| BI dashboard | 3 | Power BI or Looker Studio | Ship a dashboard that refreshes on its own |
| Statistics and EDA | 3 | Your BI tool and a notepad | Spot a real pattern and rule out a fake one |
| Python with pandas | 4 | Python and pandas | Automate a cleanup and rerun it monthly |
A suggested part-time plan, about 17 weeks or four months of steady evenings. Adjust the weeks to your own pace, the order matters more than the timing.
Skills that are not tools
The tools are only half the job. The other half is a set of skills no download gives you, and they are the ones that separate an analyst who gets promoted from one who stays stuck. The first is asking a sharp question, because a vague request like tell me about sales produces a useless answer, while a sharp one like which region lost the most repeat customers last quarter points straight at the data you need. The second is communication, because a correct finding that a busy manager cannot follow may as well not exist. You will spend real effort learning to say the one sentence that matters and cut the rest.
The third is business context, which just means understanding what the company is trying to do so your numbers point at something that matters. You practise all three the same way you practise the tools, by doing. Every time you build a chart, write the one sentence a manager would take away from it. Every time you get a question, rewrite it to be sharper before you touch the data. These habits cost nothing and compound faster than any tool, and this series builds them alongside the technical parts, most directly in Part 3 on thinking in data and Part 19 on presenting to stakeholders.
My take
If I could force one rule on every beginner, it would be this: do not open a Python tutorial until you can write a SQL join from memory. I have watched too many capable people spend their first months on the layer that feels most like real programming, then freeze in an interview that asks for a basic query. Spreadsheets and SQL are boring to brag about and lethal to skip. Get those solid, add one dashboard tool, and you are already more employable than most people who have been studying twice as long in the wrong order.
Start in a spreadsheet this week
Here is the recommendation for this part, stripped to one line. Do not buy five courses and do not install Python yet. Open a spreadsheet you already have, learn pivot tables until they feel easy, and then move to SQL. That single sequence, spreadsheet then SQL then one dashboard tool, is what makes you hireable, and you can hold the whole plan in your head from the roadmap above. Everything after it is a bonus that gets easier once the base is solid.
Your task this week: take any spreadsheet you can find, a bank statement, a list of expenses, a sports table, and build one pivot table that summarises it, then write the single sentence it tells you. That is the whole loop in miniature. In the next part we slow down and learn how to think in data, turning fuzzy requests into clear questions, metrics, and hypotheses, which is the skill that makes every tool above worth using. Keep the Data Analyst guide open as your map, and look back at Part 1 if you are still deciding which role to aim for.
References
- Download Power BI Desktop, Microsoft Learn
- Data source, Looker Studio Help, Google
- PostgreSQL Tutorial, The PostgreSQL Global Development Group
- Getting started, pandas documentation
About The Author
DrJha