Preface
Goal: A practical overview of trends, from modeling data in Excel to scripting it in Python.
I can read the data.
I know what it means.
Once, I had this eureka moment.
I want you to experience it too.
Data is nowadays language, a new form of communication.
The way we perceive data shaping how we think.
In this way, insight becomes natural outcome.
Data, Math, and the Code Behind It
I’ve always wanted to learn statistics, but I never quite knew where to begin. The breakthrough came when I was challenged to fit a curve to a data series. That one task opened the door to regression, correlation, and a whole new perspective.
Back in college, I never skipped a single statistics class. But even then, what we learned barely scratched the surface, just the basics. Meanwhile, I heard that other universities were diving into things like ANOVA. Now, two decades later, I find myself learning everything from scratch. No formal class. No course. Just curiosity.
I started seriously exploring statistics around 2017. A book helped me lay the foundation, and since then, I’ve seen tons of content, concepts in videos, fancy online calculators, and summaries in articles. But here’s the problem: very few of them come with real, practical Excel examples for everyday use.
And without doing the calculations manually, how can I be sure my results are even correct? I want to see how the math works, not just get the answer.
Even the online tools fall short. If they don’t explain the math equation behind the scenes, how am I supposed to learn from them? Books help, sure. But even if I understand the theory, how do I turn that into a script? How do I actually use this knowledge? How do I automate my job with scripting?
That’s why I’ve decided to take a new approach on learning statistics. One that’s hands-on, grounded, and script-friendly:
- Understand the math behind the method
- Reproduce it manually using Excel
- Implement it programmatically using Python
- Go further with automation, nice visualization, and combine with other technologies
Approachable Statistics
Down to Earth
I want to bring statistics down from the ivory tower, and into the kitchen table, where real people make real decisions.
Let’s make statistics less of a scary math monster, and more of a friendly tool. Something we can use even before our morning coffee kicks in.
Habit Change
How do you start solving math problem?
Back in my college days, it was simple: Grab a sheet of paper, scribble down the equation, and work it through by hand. Line by line. Thought by thought. Derive this. Simplify that. Pure muscle memory.
But times have changed. Paper is rare. Pens are even rarer. These days, I don’t even carry one in my bag. No more old-school notes on napkins. So naturally, the way I “write things down” has evolved.
What tools do we used these days?
How do we approach it now? We’ve had spreadsheets for more than two decades (since about 25 years ago), and they’ve become second nature. When intuition kicks in, I reach straight for Excel It’s quick, visual, and everywhere. Need to show someone your data model? Screenshot it. Paste the result into WhatsApp. Or better yet, just attach the whole spreadsheet file.
How do you communicate the equation itself?
Sure, Excel can display equations nicely. But if I want to share the math behind the scenes, especially in a clean, readable format, I’ll use LaTeX. LaTeX gives me both beauty and precision: equations that are easy to read by humans, and still machine-readable.
What is the final form?
It depends, on the audience, the purpose, and the medium. For something like a polynomial fit, I usually want a visual chart. One that compares the raw data to the fitted curve. So anyone can instantly see what’s going on.
This chart could be made in Excel, sure. But for more control and automation, I use Python’s matplotlib. That way, I can export an image, ready to drop into any chat app or social feed. You could use R, or any language that speaks data. The tool doesn’t matter. The insight does.
How do We Start?
Hands-on walkthrough!
We kick things off with something familiar: polynomial curve fitting.
From there, we’ll level up. Venturing into least squares, regression, and correlation. And it all starts with a good old-fashioned cheatsheet.
Here’s a breakdown of the equation flow for basic linear regression:
But we’re not stopping at linear. we also tackle polynomial regression too.
To make this journey easier, I’ve put together a beginner-friendly worksheet. Aa step-by-step guide to solving linear regression by hand, with full statistical context.
Once we’ve done the math manually (because we should understand it before automating it), we’ll switch gears and use built-in spreadsheet formulas, in Excel or LibreOffice Calc.
Then comes Python.
We’ll walk through manual calculations in code,
followed by Python’s built-in helpers.
numpy, scipy, and other statistics libraries.
Then we’ll move on to visualizing results with matplotlib.
For example: here is a confidence interval shaded around a regression line.
We’ll also cover distribution analysis, plotting density curves to understand how our data behaves. We need to learn the basic of plotting the distribution curve.
Eventually, we’ll be able to overlay statistical markers on histograms, like this:
Want a little more elegance? We’ll enhance our plots using Seaborn, And yes, there’s a JupyterLab version of every script on GitHub.
We’ll even verify our manual work using PSPPire, the open-source cousin of SPSS.
Eventually, you’ll see how this entire workflow can be replicated in other languages:
such as R, Julia, and maybe later I will continue with Typescript, and Golang.
That’s all. For further information, please contact your nearest statistician.
Table of Content
Polynomial Article Series
I divide this polynomial articles series into subseries:
-
Trend: Prediction
-
Trend: Regression
-
Trend: Enhanced
-
Trend: Language
Understanding Data
Do you speak data?
The way we understand data isn’t just about numbers. It’s how we speak to the world, how we connect ideas. It’s more than interpretation. It’s a new language, a shared space where insight emerges naturally, like words forming a story. And once we start seeing it, there’s no going back.
It’s not about mastery, it’s about becoming fluent.
Data Jokes
Philosopical thought could be hard. So, let’s lower our frequency. Relax, and here is a joke for data nerds.
A model strikes up a conversation with a mathematician and proudly says:
“Hey, I’m a model, you know!”
The mathematician replies, deadpan:
“Oh? What kind? Polynomial, Chebyshev, Legendre, or Hermite?”
She meant runway. He meant regression. Neither apologized.
What Comes Next 🤔?
The journey begins with the built-in LINEST formula in spreadsheets,
and the Polynomial library method in Python’s numpy.
So if you’re ready, let’s dive deeper: [ Trend - Built-in Method ].