<![CDATA[Mobile data has become a routine monthly expense for many university students, yet few studies have modelled this expense, and fewer still show the linear algebra behind such a model. In this study we built a model of monthly internet data package expenditure using the matrix inverse method. Fourteen students from the 2023 cohort of the Mathematics Education Study Program at Universitas Kristen Indonesia completed a short questionnaire on the quota they bought each month (GB) and the amount they paid for it (thousand IDR). We arranged these responses as a design matrix, formed the normal equations, and solved them with the inverse of XTX. After checking the arithmetic, we evaluated the model with the coefficient of determination, F- and t-tests, residual analysis, and the mean absolute percentage error (MAPE). The model obtained was Ŷ = −7.271 + 2.250X, so every extra gigabyte added roughly IDR 2,250 to a student’s monthly spending. Quota accounted for 93.6% of the variation in expenditure (F(1, 12) = 175.02, p < 0.001), and the predictions missed the actual amounts by 12.60% on average. The negative constant was not significant, which means the model should only be used between 10 and 50 GB. The 30 GB packages deviated most from the line because they cost more per gigabyte than the others. For students, the model offers a quick way to budget for data; for lecturers, it is a worked example of the matrix inverse applied to data that students produced themselves.]]>
