Automation in exam grading with Python and web hosting: data science applied to education

Abstract

Data science employs statistical methods and computational tools to automatically extract knowledge from data or information on a specific subject. In the educational context, automating exam grading enhances efficiency and accuracy, which is especially relevant in social and agrarian sciences, where a high volume of students, questions, and the need for quick grading are common. Thus, the objective of this study was to develop an automation system called AutoCorrect, utilizing Python and web hosting for the grading of academic exams. The mechanism allows for the generation of customized exams with randomized answer keys and performs automatic grading of submitted responses, significantly optimizing the evaluation process. The methodology included the use of specific Python libraries, such as Pandas, Numpy, and Matplotlib, for data processing, statistical analysis, and graphical visualization. A practical application was conducted in Agronomy and Veterinary Medicine classes, in which students took the generated exams. Student responses were directly compared to the answer key provided by instructors, resulting in the automatic generation of scores for each student based on correct and incorrect answers. Additionally, the system enabled a detailed analysis of grades through descriptive measures, such as mean, mean standard deviation, variance, and skewness, providing a clear view of class performance. Supplementary charts were generated to facilitate visualization of the grade distribution and individual performance relative to the overall average. The integration of data science with computational tools like Python and web hosting not only optimizes the grading process but also provides greater clarity and agility in analyzing student performance. The AutoCorrect system proves to be an efficient and innovative tool for various educational fields, especially in agrarian and social sciences.