Last month, I was sitting in the Kalyan railway station at 7:15 AM, coffee in one hand, my laptop in the other, scrolling through a LinkedIn post from some guy claiming he learned "15 data analysis tools" and got hired within three weeks. I believed him for exactly forty seconds.
Here's why: the week before, I'd interviewed three candidates for a junior analyst role at Morningstar. All three had certificates in Power BI, Tableau, and Python. One had even completed a three-month bootcamp. But when I asked them to walk me through how they'd approach a simple real-world problem—analyzing customer churn using messy transaction data—two of them froze. The third started talking about algorithms before understanding the business question.
That's when it clicked. The skills that get you hired in 2025 aren't the tools. They're the thinking patterns behind the tools.
I'm not saying bootcamp certificates are useless. I'm saying they're incomplete. And I've spent the last five years watching people waste time, money, and genuine enthusiasm chasing the wrong skills while companies desperately need something else entirely. This post is me cutting through the noise and telling you exactly what I see working—and what I see failing.
The Skill Nobody Talks About: Asking the Right Question First
When I joined Morningstar, I was technically decent with Excel and had decent SQL knowledge from my Economics postgrad. But I was useless at data analysis.
The first real project I got assigned: "Analyze our fund performance against benchmarks." Simple enough, right? I spent two days building queries, pulling datasets, creating visualizations in Excel. I had pivot tables. I had conditional formatting. It looked thorough.
My manager looked at it for thirty seconds and asked: "What decision should this analysis help us make?"
I had no answer.
That conversation changed everything for me. Because that's the skill nobody teaches you in data bootcamps.
The Business Questioning Framework
What actually gets you hired is this: before you touch a single dataset, you ask yourself four questions.
One: What decision are we trying to make? (Not "what data do we have"—what *decision*?) Two: Who needs this answer? Three: What would change if we knew this? Four: How confident do we need to be?
I used to skip all four and jump straight to Excel. Now I spend 40% of my analysis time just understanding the problem. And honestly? That's why I get hired to lead projects.
Here's what I see in most bootcamp graduates: they can build a dashboard in Power BI in under an hour. But ask them why we're measuring *that particular metric* and not something else, and they go quiet. That's a ₹50,000+ skill gap right there.
Why Companies Care More About This Than Your Tool Fluency
Companies don't actually care if you use Python or R. They care if you use the right one for the problem. They don't care if your dashboard is built in Tableau or Power BI. They care if it answers the question nobody else thought to ask.
The analysts who get promotions? They're the ones who walk into a meeting with data that changes how people think—not just data that answers what they asked.
The Technical Skills That Actually Matter in 2025
Okay, so questioning and thinking come first. But let's be real—you still need to *do* something with data. So what should you actually learn?
I talk to a lot of data professionals, and I've noticed something: everyone's obsessed with learning everything. Everyone wants to be the person who knows machine learning *and* dashboarding *and* SQL *and* Python *and* Tableau. You know what that person usually is? Mediocre at all five things.
Here's what I think actually works in 2025:
SQL is Not Optional Anymore (Sorry)
I used to think SQL was just for data engineers. I was wrong. Every analyst role I've seen posted in the last year requires SQL. Not "nice to have"—*requires*.
And here's the thing: it's not hard. It's actually simpler than Python. You can be genuinely useful with SQL in four weeks of consistent practice. Not "expert." Useful. Enough to pull your own data, clean it, and do exploratory analysis without bothering the data engineering team every time you have a question.
That single skill—the ability to pull data yourself—moves you from "analyst who needs support" to "analyst who can move fast." That's a ₹3–5 lakh difference in hiring bands.
Learn: SELECT, WHERE, JOIN, GROUP BY, HAVING, and window functions. That's 80% of what you'll actually use. Practice on Zerodha's public datasets or build queries against public databases. Don't do a bootcamp. Just practice.
Excel Still Matters (But Differently)
I know people who say "Excel is dead" in 2025. These people don't work in actual companies.
Every single day at Morningstar, I send someone an Excel file with calculations. Not because we're stuck in 2010—because Excel is still the fastest way to communicate numbers with non-technical people. And 90% of your job is communicating with non-technical people.
But here's the shift: in 2025, Excel matters less for raw data crunching and more for *storytelling with numbers*. You need to be good at pivot tables, of course. But more importantly, you need to know how to format a sheet so a CEO can understand your point in fifteen seconds. You need to know when to use a chart versus a table. You need to know conditional formatting, simple formulas, data validation.
The Excel skill that gets you hired: building a model that someone else can modify. Building dashboards that update automatically. Taking messy source data and turning it into something clean that tells a story. That's not what most people practice.
One Visualization Tool—Pick One and Get Really Good
Should I learn Power BI or Tableau or Looker? This is the wrong question.
Here's the right question: which one does the company I want to work for use? Learn that one. Get *really* good at it. Not "I can build a basic dashboard." I mean: you can build something that makes people stop working and actually look at it.
Most candidates learn dashboarding from tutorials. They know how to make a bar chart. They don't know how to design a dashboard that actually drives decisions. They don't know how to tell a visual story.
Spend three months with Power BI. Build 10 dashboards. Publish them somewhere. Let other people use them. Fix things based on feedback. That's worth more than "learning" four different tools.
The Soft Skills That Companies Actually Prioritize
This is the part nobody tells you because it's not sexy enough to sell a course about.
But in every interview I've done, in every promotion I've seen, in every "wow, they're really going places" conversation I've heard—it comes down to soft skills. And I mean specific ones, not just "communication" in a vague way.
Explaining Technical Concepts to Non-Technical People
This is the highest-paid skill in data analytics. Not in theory—in practice.
I know analysts who are technically brilliant but can't explain why their analysis matters to a product manager or a CFO. Those people hit a ceiling at mid-level roles and stay there.
I know analysts who are technically fine but can explain complex findings in two minutes using the right examples and analogies. Those people get promoted into leadership.
How do you build this skill? Practice. Deliberately. Take a complex analysis and explain it to someone outside your field. Your parent. Your friend who works in sales. Your partner. Can they understand *why* it matters in 30 seconds? If not, your explanation isn't good enough.
The Ability to Work with Bad Data and Still Deliver
This one surprised me. Every analyst role requires "working with data quality issues." Most people think this means "knowing how to clean data." It actually means "knowing how to deliver useful insights even when your data is messy."
Real example: last quarter, we had a dataset where 40% of customer IDs were duplicated across systems. Most analysts would have said "we can't analyze this until it's fixed." I built analysis around what we *could* trust, documented the limitations, and gave the business what they needed to make a decision. That got me noticed.
The skill: not being afraid of imperfect information. Understanding that "good enough for a decision" is different from "academically perfect." That's what actually matters in companies.
The Productivity Skill: Automation Thinking
If you're doing the same task twice, you should have automated it the first time. This mindset—automation thinking—is what separates junior analysts from senior ones.
Most people are comfortable with this concept. Almost nobody actually does it. They manually create reports. They manually update spreadsheets. They manually send emails with attachments.
Learning a little Python for automation? That's not "nice to have" in 2025. That's table stakes. You don't need to be a programmer. You need to be able to write scripts that save you five hours per week. Once you do that once, you'll do it forever.
| Skill Category | In Demand 2025 | Time to Learn (Practical Level) | Salary Impact |
|---|---|---|---|
| SQL (Core) | Essential | 4–6 weeks | ₹3–5 LPA increase |
| Excel (Advanced) | Essential | 6–8 weeks | ₹2–3 LPA increase |
| Power BI/Tableau | Essential | 8–12 weeks | ₹2–4 LPA increase |
| Python (Automation) | High | 10–14 weeks | ₹3–6 LPA increase |
| Business Questioning | Critical | Ongoing (6+ months) | ₹5–10 LPA (promotion) |
| Communication/Storytelling | Critical | Ongoing (3–6 months) | ₹4–8 LPA (promotion) |
What I Got Wrong (And What Changed)
I used to believe that more tools = more employable. I spent six months learning R, then six months learning Scala, then three months with machine learning libraries. Know what happened? I was mediocre at everything.
Then, during my MA Economics, we had this professor—Dr. Sharma—who would always say: "It's not about having answers. It's about asking better questions. Better questions lead to better decisions. Better decisions lead to better outcomes." I thought it was just philosophy class stuff.
But that's literally it. That's the entire skill set.
What surprised me most? Companies don't want a data scientist. They want someone who can turn data into decisions. That's a different skill entirely. Someone can know machine learning algorithms inside-out but be useless if they can't articulate *why* a model matters or when it should be trusted.
I'd do things differently if I started over. I'd spend the first two months just learning to ask questions properly. Then I'd learn SQL deeply. Then Excel. Then one visualization tool. Then Python for automation. And I'd spend time on communication skills—not as an afterthought, but as a core skill alongside technical ones.
The people getting hired right now? They're not the ones with the fanciest GitHub profiles. They're the ones who show up to an interview and within five minutes make it clear they can take messy, complex information and turn it into something useful.
Final Thoughts
If you're sitting where I was five years ago—either in a job that feels stuck or considering jumping into data analysis—here's what I genuinely believe will work in 2025:
Stop trying to learn everything. Pick three things: SQL, Excel, and one visualization tool. Spend three months on each. Do real projects. Build a portfolio that shows your thinking, not just your technical chops. Get comfortable with the phrase "I don't know, but here's how I'd figure it out."
That combination—technical skill plus thinking skill plus communication skill—is what actually gets you hired. And more importantly, it's what gets you promoted, respected, and paid well.
The commute from Kalyan to Mumbai is long, but it's given me time to think about what actually matters in this field. And it's not the tools. It's never been the tools.
You've got this. Start with one skill. Get really good. Then add the next one. That's the path that works.
Dattatray Dagale
Data Analyst • Blogger • Mumbai
I'm a data analyst from Kalyan, Maharashtra, working at Morningstar. I write about personal finance, career growth, and everyday life for Indian millennials — the stuff I wish someone had told me earlier.
Written by Dattatray Dagale • 31 August 2026
.gif)
.png)
.png)
0 Comments