Maths IA guide
Using Desmos, GeoGebra and Excel in your Maths IA
Technology is expected in the Maths IA: it lets you work with real data and try models you could never fit by hand. But the software doesn't earn marks — your explanation of what it did and why does. This guide shows which tool suits which job, a step-by-step regression in Desmos, and how to write about technology so it helps Criteria B and E instead of hurting them.
Which tool for which job
| Tool | Best for | Notes |
|---|---|---|
| Desmos | Fitting models to data (regression with any formula you type), comparing models on one graph, sliders to explore parameters, residual plots. | Free, in the browser; graphs export cleanly. Regression uses least squares on the formula you give it. |
| GeoGebra | Geometry and Voronoi diagrams (the Voronoi command), loci, regression in the spreadsheet view, CAS for checking algebra, 3D graphs. | Free; the CAS view is useful for checking, but your working must still be shown and explained. |
| Excel / Google Sheets | Cleaning and summarising large datasets, repeated calculations (e.g. Euler's method, loan schedules, Markov chain steps), statistics functions such as CORREL, RSQ, LINEST, CHISQ.TEST and T.TEST. | Ideal when the same calculation runs hundreds of times; show one step by hand, then the table. |
| GDC (your calculator) | Quick checks, statistical tests and regression you already know from class. | Fine for results; screenshots of a calculator screen look poor, so recreate the graph properly. |
| Python or other code | Simulations (probability, games), large computations, numerical methods. | Allowed; explain the algorithm in words and mathematics, and put the code in an appendix. |
How to fit a model and check its residuals in Desmos
- Enter your data in a table. Add a table (the + button, then “table”) and type or paste your data into columns x₁ and y₁. Paste straight from a spreadsheet if you have many rows.
- Plot it and look before you fit. Adjust the axes so all points show. Describe the shape in words and decide which family of functions the context suggests, before fitting anything.
- Type the regression. In a new line type y₁ ~ a·x₁ + b for a straight line, or y₁ ~ a·e^(b·x₁) for an exponential, or any formula you have justified. The tilde (~) tells Desmos to find the parameters by least squares.
- Read the parameters and the fit statistic. Desmos shows the fitted values of a and b and a measure of fit (R² or r). Record them with sensible rounding and interpret each parameter in context.
- Plot the residuals. Click “plot” next to the residuals (e₁) in the regression line to see them. A random scatter around zero supports the model; a pattern means the model is missing something.
- Compare a second model. Fit an alternative formula to the same table and compare the residuals and the sum of squared residuals, not just R². Say which model you choose and why.
- Export and label. Export the graph as an image, then give it a title, axis labels with units and a figure number in your document, and refer to it in the text.
Try it now: desmos.com/calculator is free in any browser.
Writing about technology so it earns marks
- Say which software you used and for what, once, clearly: “The exponential model was fitted in Desmos by least-squares regression.”
- Explain what the tool did mathematically. For regression, that it chose the parameters that minimise the sum of squared residuals.
- Interpret every number the tool gives you. An R² you never comment on earns nothing.
- Show at least one calculation by hand where it matters (one Euler step, one expected frequency), then let the tool do the rest.
- Recreate or tidy graphs so they meet the labelling rules; don't paste a cluttered screenshot.
- Acknowledge the software in your bibliography like any other source.
Frequently asked questions
Is it allowed to use Desmos or GeoGebra in the Maths IA?
Yes. Technology is expected. Acknowledge the software you used and explain what it did.
Does using technology lower my Criterion E mark?
Not if you explain and interpret what it did. It lowers the mark when output replaces understanding — for example a regression equation or a p-value quoted with no explanation of where it came from or what it means.
Should I use R² or r?
Use r (Pearson's correlation coefficient) for linear relationships and R² (the coefficient of determination) to compare how well different models fit, and explain what each measures. Neither alone proves a model is right: look at the residuals too.
Can I include code in my Maths IA?
Yes, in an appendix. In the body, describe the algorithm and the mathematics it implements, and interpret the output.
See technology written up well
Every exemplar in the IA package uses Desmos, GeoGebra or a spreadsheet, with the examiner's note on how the technology is explained. €39 once, 12 months' access, 14-day money-back guarantee. It helps you write your own IA; it never writes it for you.