EXCEL to SQL: My Next Step as a Data Analyst
A few weeks ago, when spending a long practice session cleaning dirty data, creating various formulas, and building a dashboard, I understood that everything I had created until that moment was contained within a single spreadsheet.
The next skill on my list wasn’t another Excel trick. It was SQL.
This is a short account of that transition,where my data analyst training started, why Excel eventually asks you to leave the building, and what the first weeks of actually learning SQL have looked like.
Where I started: Excel as a training ground
My data analyst training started inside Excel, and that’s exactly where I’ve spent real time sharpening the fundamentals:
- Cleaning and standardizing messy, inconsistent data before it’s usable
- Writing formulas and lookups that pull information across sheets
- Summarizing large datasets into pivot tables a manager can actually read
- Visualizing results and flagging outliers so they’re obvious at a glance
- Building small working data models out of several related tables
As a result, I completed several capstone projects, including analyzing of six tables with feedback from a university and a dashboard for a travel agency operations. Here you see a chart from one of the little exercise, it provides a simple comparison that is sorted and colorized in order to notice the outlier without numbers’ reading.
Where Excel starts to strain
Excel is able to do much more than most people think it could do. So, in my practice capstone, I managed to join and summarise six interconnected tables (15k of booking records, 10k of clients and 12k of their opinions) into ranked pivot tables fully manually. It became possible, but at the same time showed a quite obvious limitation: each of these tables was a named range of data in one file, each of these relationships should be established manually, and any change of the question means adding a new formula to that file.
In Excel, I learned how to ask correct questions to my data. And SQL is teaching me how to ask them in mass quantities.
The pivot table below, taken from this very capstone, shows a lot of manual effort needed to establish relationships between many sheets to calculate a number of classes taught by each professor. The result looks neat, but the process of reaching the desired outcome took more effort than I would like.
Transition to SQL
I joined Learn SQL on Scrimba, an entirely free introductory course that focuses on short exercises rather than watching video tutorials. The learner receives instructions, creates a valid query in SQL against a real table, and executes it right away.
Get Fredrick Bundi’s stories in your inbox
Join Medium for free to get updates from this writer.
These early exercises feel familiar because the tasks are filtering a table with cars of a certain color (red), year within the sixties, price between two values and return just the needed columns. This practice reminds me of Excel logic of filtering and summarizing information, just using SQL as a language made for that rather than attached to a spreadsheet application.
What Fundamentals of SQL has taught me so far
From ten lessons in the module Fundamentals of SQL, several important things were revealed for me:
SELECTpulls back exactly the columns I ask for, nothing more, instead of building a new lookup every time I need a different viewWHEREcombined withLIKEand wildcards replaces a pile of conditional formulas stacked on top of each otherBETWEENhandles numeric and date ranges in one clean clause, instead of combining several conditions just to check a rangeORDER BYdoes instantly what used to mean selecting a range and sorting it by hand
Nothing surprising there since this logic is close to what was used before in Excel, just written in one sentence, which is executed immediately on the tables that are able to store much more data compared to spreadsheets. Approximate example of the query I create now:
SELECT product_name, category, price
FROM inventory
WHERE category = 'Electronics'
AND price BETWEEN 50 AND 200
ORDER BY price DESC;
Excel Vs SQL
I will keep using Excel since it is good for creating dashboards navigated by users, where conditional formatting shows potential problems at once and when a one-time job takes ten minutes. SQL is useful for a completely different set of problems, when there is a database in the very beginning of the data work, and the task is to pick needed rows from it, not to reshape existing data.
Two tools, same analyst, different jobs.
What’s next
At this moment, the plan is to complete the Fundamentals of SQL module, then take Creating and Joining Tables and finally Advanced Logic and Conditions. After getting a certificate the plan is to look through all the catalog and figure out next courses, and the choice is wider than I thought, there are JavaScript, React and even coding together with an AI assistant course there.
A few things I’d tell someone starting out
- Make sure you master Excel well first. Cleaning data, using formulas, and manually summarizing the data develops basic skills of data manipulation that any further tool presupposes.
- Do not overlook the tedious work of data cleaning. It takes up most of the analysis work in the real world: trimming spaces, fixing data types, removing duplicates.
- Treat your mistakes as a part of the process, not the sign that you have chosen an inappropriate area of study. In the beginning of the formula learning process, your mistakes will be expensive in terms of time, but you will gain much more from them than from any tutorial.
- It is time to learn queries when a spreadsheet becomes limiting to you.
Try it yourself
If any of this sounds like where you are, stuck at the edge of what a spreadsheet can comfortably do, SQL is a short, practical next step rather than a wall to climb. I started with Learn SQL on Scrimba, free, beginner-friendly, and built around actually writing queries instead of watching someone else write them.
If you give it a try, I’d genuinely like to hear where you’re starting from. Drop a comment, I’m curious what everyone else’s “Excel ceiling” moment looked like.
