Mobasshir Hasan

Empowering Education Through Internet Access: The Impact of Digital Devices Worldwide Mobasshir Hasan, MBA’24 Babson College Background Literacy is a fundamental element of social and economic progress. Despite its importance, many areas around the world continue to face low literacy rates. The lack of access to educational resources, the Internet, and digital devices plays a significant role in this ongoing issue. This study focuses on utilizing data related to Internet accessibility, mobile device ownership, and other socioeconomic factors to pinpoint the main drivers that can improve literacy rates. By understanding these key elements, this study aims to offer practical recommendations for policymakers to enhance educational access and opportunities, ultimately fostering a more educated and connected society. Methodology To achieve the objective, this study integrates nine datasets, encompassing Literacy Rates, Households with Internet Access, Individuals Using the Internet, Mobile Broadband Traffic, Population Coverage by Mobile Network Technology, Fixed Broadband Traffic, Individuals Who Own a Mobile Cellular Telephone, Fixed Broadband Internet Basket Costs, and Mobile Data and Voice Low Consumption Basket Costs. This study standardizes column names and merged datasets using country and year as keys, addressing missing values through interpolation and placeholders. We create a binary target variable, EducationBinary, based on a literacy rate threshold of 60%. Interaction terms and polynomial features, such as InternetAccessPopulationInteraction, MobileTrafficPopulationInteraction, InternetUsageSquared, and MobileOwnershipSquared, were added to capture complex relationships. The dataset was split into training and testing sets, and we developed and evaluated Logistic Regression and Random Forest models to identify significant predictors of education levels. To understand the factors influencing educational outcomes, we used Logistic Regression and Random Forest models. Logistic Regression showed that countries like Barbados, Gibraltar, and Cuba, with high Internet Usage and Low Broadband Costs, significantly predict higher education levels. This model helped us pinpoint key factors driving better educational outcomes, offering guidance for improving literacy rates by focusing on enhancing these features in other regions. The Random Forest model highlighted Internet Usage as the most critical feature, strongly linked to higher literacy rates. Lower Fixed Broadband Costs and broader Population Coverage by mobile networks were also important, supporting better educational outcomes. The model's accuracy was 59.37%, with a sensitivity of 0.7374 and specificity of 0.4161, indicating its effectiveness in identifying countries with high Literacy Rates. Overall, these findings suggest that increasing internet penetration, reducing broadband costs, and expanding mobile network coverage are effective strategies for enhancing literacy rates globally. Analysis The heatmap of the correlation matrix revealed strong correlations between Fixed Broadband Cost and Mobile Data Cost, and between Population Coverage and Internet Usage. This suggests that higher internet penetration and lower costs are likely to have better educational outcomes The histogram shows the distribution of Internet Usage by Education Levels. Countries with higher Education Levels (right) exhibit higher Internet Usage, suggesting a strong link between Internet access and better educational outcomes. The scatter plot with regression line for Internet Usage vs. Mobile Ownership showed that countries with higher internet usage also had higher mobile ownership, indicating a positive relationship. Figure 7. Variable Importance based on Random Forest Model Figure 6. Confusion Matrix from Random Forest Model Conclusion Figure 1. Heatmap of Feature Correlations: Identifying Key Relationships Influencing Educational Outcomes Figure 3. Scatter Plot with Regression Line: Figure 2 Hisogram distribution of Internet Internet Usage vs. Mobile Ownership and Its Usage by Education Level Impact on Education This study underscores the pivotal role of internet accessibility and affordability in driving global educational outcomes. By integrating and analyzing multiple datasets, we identified that higher internet usage, lower broadband costs, and extensive mobile network coverage are strongly correlated with increased literacy rates. The machine learning models, particularly the Random Forest classifier, highlighted these factors as crucial determinants of educational success. These insights provide actionable intelligence for policymakers and stakeholders aiming to improve literacy rates through targeted investments in internet infrastructure and affordability, along with educational support such as teaching, digital education software, and hardware. Ultimately, this research demonstrates the transformative potential of digital connectivity in fostering educational development and closing the global literacy gap. Future work Further research can incorporate additional data sources, such as economic indicators, and conduct temporal and region-specific analyses. Utilizing advanced modeling techniques like deep learning can improve predictive accuracy. Longitudinal studies can provide insights into how changes in digital access impact education levels over time. Additionally, exploring the interplay between education and other domains, like health and employment, will offer more actionable insights for policymakers. Density plots for Mobile Ownership and Internet Usage by EducationBinary classes demonstrated that countries with higher education levels had higher densities of mobile ownership and internet usage, underscoring the importance of internet access for education. References Data Sources: 1.UNICEF: "Education and Literacy Data (https://data.unicef.org/resources/dataset/education-literacy-da). 2.World Bank: Datasets on internet usage, mobile broadband traffic, population coverage, and costs (https://data.worldbank.org). Figure 4. Density Plot of Mobile Ownership by Education Level Figure 5. Density Plot of Internet Usage by Education Level Other References: 1.James, Gareth, et al. "An Introduction to Statistical Learning: with Applications in R." Springer, 2013. 2.Goodfellow, Ian, et al. "Deep Learning." MIT Press, 2016.