IA modelling · interactive helper

Model my data: the play-by-play with your numbers

Paste your x and y values, choose a model and see every step of the method with your own numbers: parameters, working, residuals, R², SSR and a graph. Then compare several models, including a test on points held back.

Private by design: the calculations run in your browser; nothing you type is uploaded, saved or stored.

1. Your data x and y values, one pair per line (paste two columns from a spreadsheet, or type e.g. 1.5, 3.2)

A header line is skipped. Decimal points, not commas, unless the columns are separated by tabs or semicolons. Up to 500 points.

x label (optional)y label (optional)

Or load example data

2. The model Model

First point (number in your list)Second point

Point numbers to useSecond point numberThird point number

Blank: first, middle and last.

Turning point = point numberor hand kSecond point number

Period from the context (optional) Keep the period fixed when refining

Vertical asymptote x = h (from the context; 0 if none)

Horizontal asymptote y = c (from the context; 0 if none) Also refine by least squares on the original data …and let c vary

Also refine by least squares on the original data

Starting value for L (optional: the ceiling from the context)

Split after x =Second piece

Hold out the last points when comparing

How the helper works

Straight lines and anything that can be made straight (logarithmic, reciprocal, exponential, power) are fitted by the least-squares formulas, shown with the Σ table — the same as a GDC.
Quadratics through three points are solved by elimination; quadratic and cubic regressions by least squares.
Sine and logistic models start from the by-hand estimates on the worked pages and are refined by the Levenberg–Marquardt method (an improved Gauss–Newton method), which is how calculators fit them too.
Every model reports residuals, SSR, R² and RMSE in the original units, so different models can be compared fairly.

Each model has a worked page explaining the method in full: straight line, quadratic, sinusoidal, logarithmic, reciprocal, exponential, power, logistic, cubic and polynomials, piecewise, and choosing and comparing models.

Frequently asked questions

Can I use this helper for my IA?

Use it to learn the method and to check your own working. Your IA must show your own working and explanations; acknowledge the helper if you used it, as your school's policy requires, and never paste its output into your IA.

Is my data uploaded or saved?

No. Everything is calculated in your browser by a small script; nothing is sent to a server, and nothing is stored — reload the page and it is gone.

Why do my calculator's numbers differ slightly?

For exponential and power models a GDC fits a straight line to the logs, while direct least squares (Desmos's default) minimises errors in y itself; the helper shows both. For sine and logistic fits, iterative methods can stop at slightly different values or give an equivalent form.