Research & Projects
Research & Projects
This report examines exploratory data analysis (EDA) and its impact on HR datasets. It also explores integrating machine learning for deeper insights, reflecting the evolving HR analytics landscape.
The primary objectives of the project were to conduct comprehensive HR data analysis to understand workforce dynamics, departmental structures, and demographics. The initial situation involved a lack of actionable insights into HR management. We sought to change this by leveraging data-driven methodologies to optimize various aspects of HR, including workforce optimization, attrition prediction, and performance evaluation. Essential considerations throughout the process included data accuracy, privacy, and alignment with organizational goals.
The findings revealed significant insights into employee behavior and organizational dynamics. We successfully optimized recruitment strategies, identified potential causes of attrition, and enhanced performance evaluation processes. Unexpectedly, we discovered correlations between tenure, project involvement, and attrition rates. The project's results will inform future HR strategies and decision-making, ensuring alignment with organizational objectives and fostering data-driven HR management practices.
This report delves into the application of graph neural network models for fake news detection on Twitter. By leveraging sophisticated machine learning techniques, the project aims to classify news propagation graphs as either real or fake news, offering a proactive approach to combat the spread of misinformation. The study underscores the significance of graph analysis in understanding the intricate relationships between users and news articles, highlighting the potential for advanced technologies to contribute to the fight against fake news and disinformation in social media ecosystems.
The project aimed to develop a graph neural network model to classify news propagation graphs on Twitter as either fake or real news. The goal was to mitigate the harmful consequences of fake news by providing a tool for early detection and intervention. By analyzing graph structures and user interactions, the project sought to alert users to potentially misleading information and prompt fact-checking before sharing. Essential considerations included leveraging graph analysis techniques, understanding user demographics, source credibility, and language use to combat the spread of misinformation effectively.
The findings demonstrated the effectiveness of using a graph neural network for fake news detection, achieving a high accuracy rate of 97%. This approach outperformed traditional natural language processing models, highlighting the potential of graph convolutional networks and attention-based graph neural networks for this task. The project's results suggest that leveraging graph structures and complex interactions between entities can significantly enhance the accuracy and robustness of fake news detection. The success of the model implies promising prospects for future applications of graph neural networks in combating misinformation on social media platforms.
The project analyzed US social media trends, revealing age and gender disparities, rapid expansion rates, and implications for societal behavior. Insights include steep age gradients on platforms like Snapchat, gender distinctions, and the impact on communication technology and mental health.
The primary objective of the project was to analyze social media usage patterns and demographics in the US. The initial situation involved a lack of comprehensive understanding of social media trends and their implications. We sought to change this by exploring demographic trends, gender distinctions, and the speed of social media expansion. Essential considerations throughout the process included data accuracy, interpretation of trends, and implications for societal behavior.
The findings revealed significant insights into social media usage across different demographic groups and its rapid expansion. We observed steep age gradients for platforms like Snapchat and Instagram, with notable gender disparities on various platforms. Furthermore, we examined the speed of social media's expansion, highlighting its impact on communication technology and societal behavior. The project's results will contribute to a better understanding of social media's influence on society and its implications for mental health and communication dynamics.