visual math studio

Turning Formulas into
a Visual Journey of Reasoning

This site is no longer just an EML showcase — it’s a gallery of mathematical intuition: Every exhibit starts with a question, explains the logic with images, then grounds the formula with a hands-on example.

Enter the Gallery Open Lab

exhibit map

Every Formula Follows the Same Learning Structure

I’ve organized the content into four exhibits. You can enter through any of them, but each follows the same rhythm: question, intuition, formula, example.

Local prediction

Taylor Series

Given only information near one point, can we guess the entire curve?Height gives the starting point, slope the direction, second derivative the bend; higher derivatives refine the approximation.

f(x)=\sum_{k=0}^{n}\frac{f^{(k)}(a)}{k!}(x-a)^k+R_n(x)

Instant rate

Derivative and Tangent

How does the average rate of change become the instantaneous speed at a point?Move the second point closer and closer; the secant slope approaches the tangent slope.

f'(a)=\lim_{h\to 0}\frac{f(a+h)-f(a)}{h}

Shape correction

Square Root Iteration

How can we compute √51 to high accuracy in seconds?Take a rectangle of fixed area N, average its two sides, and the rectangle gets squeezed toward a square.

x_{n+1}=\frac{1}{2}\left(x_n+\frac{N}{x_n}\right)

Boundary thinking

Squeeze Theorem

When a function is hard to pin down, can we sandwich it between two simpler functions?Upper and lower bounds converge to the same limit; no matter how wild the middle function is, it must converge too.

g(x)\le f(x)\le h(x),\quad \lim g=\lim h=L\Rightarrow \lim f=L

Generative grammar

EML Operator

Can a single operation generate a whole family of complex functions, like a grammar?Combine exp and ln into one binary operator, then build expression trees by recursive composition.

\text{eml}(x,y)=e^x-\ln(y)

exhibit 01 / Taylor

Taylor Series: The Curve’s “Local Fingerprint”

Taylor series isn’t about memorizing a chain of derivatives — it’s about asking: If I stand at x=0, knowing only the curve’s height, direction, and bend, how far can I predict?

Heightf(0) tells where the curve starts.

Slopef′(0) tells which way it heads as it leaves the point.

Curvaturef″(0) tells it’s not a straight line — it begins to bend.

Remainder termR_n(x) tells you “how much is left” — that’s where reliability comes from.

example / e^x at 0

e^x \approx 1+x+\frac{x^2}{2}+\frac{x^3}{6}

3rd-order correction

example / sin x

f'(a)=\lim_{h\to 0}\frac{f(a+h)-f(a)}{h}

h = 1.20

exhibit 02 / derivative

Derivative: From “Two-Point Change” to “Instantaneous Speed”

The derivative captures the first‑order information of the Taylor series. Take two points to get the secant slope, then move the second point closer to the first — the limit that remains is the tangent slope.

Average changeThe height difference between two points divided by the horizontal distance.

Shorter distanceMake h smaller, and the secant looks more and more like the tangent.

Limit remainsAs h → 0, you get the instantaneous rate of change at that point.

Area N

x_{n+1}=\frac{1}{2}\left(x_n+\frac{51}{x_n}\right)

Current estimate7.14142843True √51 = 7.14142843, squared error 2.13e-14

exhibit 02 / Newton

“The Squeeze”: Fix the Area, Squeeze the Rectangle into a Square

The method in the video is essentially Newton’s iteration (Heron’s method). It’s not a magical trick — it’s a geometric self‑correction: guessing one side too small makes the other too large; averaging them pulls both closer.

Round 07.000000 and 7.285714

After averaging we get 7.14285714

Round 17.142857 and 7.140000

After averaging we get 7.14142857

Round 27.141429 and 7.141428

After averaging we get 7.14142843

Reminder: the square root of 49 is 7; the square of 49 is 2401.

exhibit 03 / squeeze theorem

Squeeze Theorem: If the Middle Is Unclear, Look at the Sides

Some functions oscillate wildly near a point, making the limit hard to see directly. The Squeeze Theorem gives a strategy: don’t wrestle with the messy middle — use two simpler boundaries to pinch it.

Upper boundFind h(x) that f(x) never exceeds.

Lower boundFind g(x) that f(x) never falls below.

PinchingIf both g and h squeeze toward the same L, the middle has no choice but to go there.

example

-|x|\le x\sin\frac{1}{x}\le |x|

Limit is 0

\text{eml}(x,y)=e^x-\ln(y)

exhibit 04 / EML

EML: Seeing Functions as a Generative Tree

This exhibit keeps the most distinctive part of the original site: how a single binary operator recursively generates complex expressions. It fits naturally into this summary site as an example of a “generative grammar for math.”

Define the building blockFirst specify the single operation eml(x,y).

Substitute inputsx, constants, or another eml can all become child nodes.

Tree compositionComplex functions arise from nesting these nodes together.

Enter the Math Formula Lab

how to read

How to Read This as a Beginner: Don’t Memorize, Understand the Moves

The point of math visualization isn’t to make formulas look nicer — it’s to expose the “why” behind them. This site can later expand with exhibits on calculus, probability, linear algebra, and number theory.

Ask a Question

Formulas don't come from nowhere; they answer a very specific problem.

Look at Geometry

Translate symbols into lengths, areas, slopes, boundaries, or trees, and abstraction suddenly feels concrete.

Try a Number

Run a small example like 51, 0, sin x — it sticks better than memorizing definitions.

Watch the Error

Real understanding isn’t just "getting the answer" but knowing why you’re getting closer and closer.

Turning Formulas intoa Visual Journey of Reasoning