Supervised Project · SCE · Computer Science · Beer Sheva

Platform for Identifying Causes of Mathematical Errors

Calc Master is an interactive calculus-learning platform designed to explain why an answer is wrong, track recurring weaknesses, and direct students toward more targeted practice.

2024–2025Shahar Levi · Lee WilnerCalculus educationError-specific feedbackSymPyAdaptive practice
4core calculus topics
4difficulty modes
15questions in mixed-test mode
1feedback explanation per selected error
Educational problem

A wrong answer is more useful when the system can explain the mistake behind it.

The project focuses on calculus practice for limits, derivatives, integrals, and critical points. Rather than showing only whether an answer is correct, the platform maps each distractor to a specific explanation so the student receives feedback tied to the particular error they made.

The broader objective is to combine immediate feedback with performance tracking so students can identify recurring weaknesses and practice those areas more deliberately.

Core learning features

  • Dynamic question generation
  • Multiple-choice exercises
  • Error-specific explanations
  • Topic-level performance tracking
  • Difficulty selection
  • Personalized and mixed-topic practice
System workflow

Generate, answer, diagnose, track, and practice again.

The implemented system uses SymPy for symbolic question generation and MathJax for mathematical rendering. Student responses feed into topic-level statistics and a dashboard that surfaces progress, best attempts, and weaker areas. A personalized quiz mode selects practice based on the student’s lowest-performing topics.

Calc Master workflow from choosing a calculus topic through dynamic question generation, multiple-choice answers, error-specific feedback, performance tracking and personalized practice.
The platform treats the chosen wrong answer as diagnostic information rather than only as a failed attempt.
Implemented functionality

A working platform rather than only a pedagogical concept.

Dynamic questions

Dedicated generators create exercises across limits, derivatives, integrals, and critical points, with multiple difficulty levels.

Targeted feedback

Each wrong option can carry its own explanation, allowing the system to respond to the specific misconception implied by the selected distractor.

Dashboard analytics

The interface tracks per-topic performance, average and best scores, and progress across attempts.

Personalized practice

The system includes focused-topic practice and a mixed 15-question mode, with basic personalization based on the student’s weaker topics.

Current limitations

The platform diagnoses selected errors, but it does not yet understand arbitrary mathematical input.

The current version is limited to multiple-choice responses. Personalization is also deliberately simple: the system prioritizes topics with the lowest score rather than learning a richer student model.

The report identifies open-form symbolic input, broader question types, deeper teacher analytics, and a mobile-friendly application as natural extensions.

Technical stack highlighted in the report

  • Python
  • Flask
  • SymPy
  • MathJax
  • HTML / CSS / JavaScript
  • Statistical progress tracking
Educational interpretation

Use assessment data to guide the next exercise, not just to report a score.

The project’s central design choice is to treat mistakes as structured data. The selected distractor, topic, difficulty, and accumulated performance history can all contribute to a more informative feedback loop. Even with a relatively simple personalization rule, the system moves from static practice toward targeted self-paced learning.

Project materials

Final report.

The student final report for this project, as a PDF.

Supervised project

From calculus exercises to a feedback-driven learning system.

Calc Master combines symbolic question generation, error-aware explanations, and student progress analytics in a deployed interactive platform.