Does revision time actually predict exam performance?
1d ago34 views
Learn with data. Think clearly with AI.
Give students a practical introduction to data and AI through questions they can investigate. Quill turns analysis into interactive, evidence-led reports where learners can test inputs, inspect how an AI-supported answer was produced and explain what the evidence really shows.
Works with your MCP-capable agent
Year 12 Data Science, investigations.
Continue reading
Browse library →Air quality around the school gates, hour by hour
2d ago27 views
Individually unpredictable, collectively obedient: 290,430 earthquakes
3d ago19 views
Reaction times across 1,200 classroom trials
4d ago12 views
What 154 years of football results teach about bias
5d ago9 views
LatestStories
AllFollowingNewCurated
Does revision time actually predict exam performance?
A class investigation into 480 anonymised study logs, testing a claim students expected to hold and finding where it breaks down.
Year 12Report
Air quality around the school gates, hour by hour
Sensor readings across a term, with the method, sources and limitations published beside every chart.
STEM clubReport
Reaction times across 1,200 classroom trials
Students changed the grouping, narrowed the age range and watched which conclusions survived and which quietly collapsed.
PhysicsReport
What 154 years of football results teach about bias
A worked example for teaching sampling, confounders and the difference between a pattern and an explanation.
MathsReport
Individually unpredictable, collectively obedient
290,430 earthquakes, used to introduce distributions, log scales and the limits of prediction.
GeographyReport
Recorded in Quill · Library → report → live results
A finished answer hides the most valuable thinking.
Students can ask an AI for an answer in seconds. The harder and more useful skills are knowing what to ask, checking where the data came from, spotting weak assumptions and deciding whether a conclusion is justified. Those skills are difficult to teach when the working disappears inside a chat or a static chart.
Introduce AI through a real investigation.
Start with a question students care about and a dataset they can understand. Use an agent to help analyse it, then review the result in Quill. Learners see AI as part of a practical workflow, with human judgement before anything is shared.
Teach students to question an AI-produced answer.
Keep the claim, evidence, source and method together. Ask learners to check whether the analysis answers the original question, identify what is missing and challenge conclusions that go beyond the data.
Let students test what changes.
Add controls that allow learners to alter inputs, compare groups or narrow a period. They can see the result respond and learn which conclusions hold, which depend on an assumption and which need another question.
Make data literacy visible.
Publish definitions, data sources, analytical choices and limitations beside the charts. Students learn to read a visual, trace a number and separate what the data shows from how someone has interpreted it.
Turn technical work into a clear explanation.
Give learners a format for presenting a claim, supporting it with visual evidence and explaining uncertainty. Share one link for a lesson, independent investigation or project presentation.
Build a reusable library of learning.
Keep investigations, worked examples and student-facing reports organised by subject. Revisit an earlier model, compare approaches and build a curriculum that grows instead of scattering work across files.
01A question worth investigating02AI-supported analysis03Visible sources and methods04A conclusion students can defend
Ask. Test. Check. Then explain.
Start with a question and a real dataset. Let an agent support the analysis, then make the working available for students to inspect. Learners can explore the evidence, change an input and explain whether the conclusion stands up.
01 / Investigate
Start with a question.
Use a real dataset and ask students to predict what they expect to find before they explore it.
02 / Explore
Make the evidence move.
Let learners change inputs, compare groups and see how each choice affects the result.
03 / Explain
Show the reasoning.
Keep the claim, visual evidence, method and limitations together in one report they can revisit.
“What did the AI do, what does the evidence support, and where is human judgement still needed?”
Quill / STEM learning — a practical foundation for learning with AI and data.
Teach students to use AI without outsourcing their thinking.
Choose one STEM topic where students usually receive a finished answer. Give them the data, let an agent support the investigation and use Quill to expose the evidence and method. Then ask students what they trust, what they would change and what the result allows them to claim.