I Completed Five Years in Mathematical Communication: 5 Lessons That Changed the Way I Work
I started my first full-time job as a Senior Data Analyst at a leading health insurance company right after finishing my schooling, bringing a strong foundation in math and business.
Five years later, I have now worked in a variety of analytical skills, including reporting, data visualization, stakeholder management, business strategy discussions, and, most recently, AI-assisted development. These five years have taught me lessons that have less to do with specific tools and more to do with understanding people, decisions, and results.
My journey into data and analytics began seven years ago, when I started grad school to study math and business. Alongside my studies, I first trained as an R&D Intern working with data sources and developing BI solutions. Then came my Data Science internship where the code grew more complex, the data messier, and the dashboards needed to meet higher standards. That experience was the cornerstone of my current success.
I realized that being someone who writes code in Python or crunches numbers is not enough. I have to be a strategic problem solver.
Reflecting on half a decade in the role, from Data Analyst to Senior Analytics Consultant, I have seen three major shifts:
- Mathematics has become more business driven than technical
- Storytelling is more important than reporting
- AI is reshaping the meaning of “technical skills”
Looking back on my time as a Math Consultant, I want to share five lessons that have changed the way I look at my work and can help anyone working in math.
1. Telling stories about data is more important than the data itself
As you grow in your career, finding yourself more often in decision rooms, you quickly realize that data alone rarely makes an impact. How that data is transmitted and used is what really affects the results.
In my experienceworking with stakeholders with varying levels of technical experience, the stakeholder may not remember the regression model they built or understand the accuracy of the model, but they certainly remember the story that helped them make the decision. The value of data is not in its mere existence but in its ability to be understood, trusted, and acted upon.
At a conference a few years ago, a speaker shared that narrative makes data more memorable than numbers alone, and that stuck with me. Since then, I have approached most of my analysis with three simple questions:
- What's going on?
- Why does it matter?
- What should happen next?
In my role as a statistical consultant, my job is not only to deliver relevant information; my job is to reduce uncertainty so that my stakeholders can work with confidence.
Data enables that process, but storytelling eliminates it.
That said, since AI becomes the “first analyst” before you touch the data, here's my caveat: telling a story doesn't mean molding the truth to fit the narrative. AI can generate compelling stories much better than a spreadsheet, but it can also present assumptions or missing numbers.
Storytelling may be more powerful than the data itself, but its power depends entirely on the integrity of the data behind it.
2. The hardest part of math is not the analysis. It asks better questions.
I was taught in graduate school that as an analyst, we should be curious. Because curiosity helps us find patterns and make sense of data. But over time, I realized that it's not just curiosity or the data itself that gives us great insight. These are the questions we ask.
You can have the cleanest data sets and the most advanced tools, but without the right questions, your analysis will drift aimlessly.
For my team of business consultants, I recently did an analytics bootcamp to teach them the basics of data and analytics. In the second week of the program, I was asked: “I can learn the tools, but how do I learn the questions I should ask as an analyst?” That was a very relevant question because when I started, I didn't have a playbook. I was always unsure what to ask the participants, what methods to use, or how to know if I found something meaningful. My goal with the bootcamp was to answer that question.
Over time, I learned that the best questions come from working closely with subject matter experts (SMEs) and unpacking the problem statement with them. These conversations reveal insights and lead your way when you have to dig deeper, which also reinforces the importance of building a strong network if an SME is unavailable.
Your takeaway is one-liner: start with curiosity, then use critical thinking. Don't jump straight to the data.
Pause and ask what is really going on, then layer your thinking on why, what, who, and when.
3. Knowing when to keep digging and when to stop
In the first few years, I truly believed that if I wanted to be a good analyst, I should not stop at the first answer. I have to collect more, filter more, ask more. That attitude helped me, until it didn't happen.
I once worked on an effort to create a service intensity report, analyzing clients who needed additional support, additional costs to the organization, and identifying what was driving service intensity. The data was incomplete and inconsistent from the beginning. However, instead of pressure testing whether it could support the project's purpose, I kept pushing forward by pulling in more data sets, testing hypotheses, and chasing the mystery that turned into noise. After about five weeks of trying to force the data to work, I finally told my boss that we can't continue.
That experience taught me one of the most important lessons I now share with every young analyst I mentor: more digging doesn't always mean more value. Somewhere, you go from getting information to wasting time getting information that no one asked for.
So now, before I go down the rabbit hole, I ask: if I find something here, will it change what I do next? If the answer is no, that's my cue to take a second look or give up, write what I have, and move on.
4. Managing expectations is part of the job done
No one tells you this in grad school, but a big part of being a successful math consultant has less to do with math and more to do with managing people's expectations of you, your data, and your time.
At first, I took all questions at face value. If a stakeholder wanted a dashboard “for the future,” I would lose sleep to make it happen, often at the cost of accuracy. It took me a while to learn that just because I can, doesn't mean I should. The real task is to have a conversation about the question: what really drives the request, what decision supports it, and what is realistic given the data we have.
A few things I do now almost automatically:
- Flag data limits in advance
- Again, ask in my words, so that the misunderstanding can arise in advance
- Talk about progress in small amounts, rather than blacking out and re-emerging with a finished product
Managing expectations doesn't mean saying no. I've learned to set healthy boundaries with stakeholders, to be honest about everything, so that trust doesn't break later.
5. AI is changing what I think “technical ability” means
When I started, being an expert meant writing efficient SQL, building clean pipelines in Python, and knowing your BI tool well enough to make it tell a story. Today, AI can write that query, write that pipeline, and suggest a type of chart before I finish typing the query. Those skills are still important, but the work has quietly changed under us.
With all the hype surrounding what AI can and can't do, the real skill of the technology now lives not in the production of work, but in judgment. I wrote a blog post recently about how metacognitive control is the most important AI skill that no one is talking about—how we need to adjust our thinking as AI takes over most of the work.
I'm sure we've all caught an AI-generated analysis confidently stating missing numbers, or recommendations that sound sharp but miss the point any analyst with six months to catch would have caught on quickly. Being “technical” today is no longer limited to coding, cleaning and converting to build a data pipeline or writing project briefs. You need to understand the data well enough to know when the AI response is subtly wrong in the first place.
From 2025, with the advent of AI, I stopped measuring my technological development by the tools I know, and started measuring it by being able to accurately assess what those tools produce.
Appreciation is a skill. Validation is a skill. Knowing when to trust the machine and when to trust your own judgment—that may be the greatest technical skill of all.
Looking Back, Looking Forward
Over the past five years, the tools I use for analysis and reporting have changed more than I expected, and I've improved more than ever. Yet my questions for any math project haven't moved that much: What happened? Why does it matter? What should we do next? Can I trust this? Should I keep digging, or should I stop?
In closing, if I had to leave one thought for anyone just starting out: data will continue to grow, tools will continue to get smarter, and AI can do more for you—but the job has always been, and always will be, about helping people make better decisions with more confidence.
…………………
That's it from my end of this blog post. Thanks for reading! I hope you found it an interesting read!
Rashi is a data wiz from Chicago who loves analyzing data and creating data stories to communicate insights. He is a full-time healthcare analytics consultant and likes to blog about data on the weekends over a cup of coffee.
