Challenges in Analyzing Unstructured Learner Generated Qualitative Big Data

Abstract

As the amount of educational data grows a necessity is occurring to define characteristics of Unstructured Learner Generated Qualitative Big Data (ULGQBD) and develop a conceptual framework to analyze these data sets. According to the author there are three main challenges in analyzing ULGQBD. First one is related to qualitative big data itself. There is an ambiguity of definition and methodology of analyzing ULGQBD. The answer of the question How much amount of qualitative data is big? is blurry. Therefore there is no a simple straightforward specific methodology to overcome the struggles originated from seeking the meaning within such a big amount of unstructured data. Second problem is absence of conceptual frameworks for analyzing ULGQBD. Third problem is related to Qualitative Data Analysis (QDA) software. Traditional QDA software is not ready to big amount of data and cloud based text analytics tools do not have enough language support for stemming for agglutinative languages such as Turkish. In this paper author shares her experience of analyzing ULGQBD. The data set which she worked on contained 13000 responses to an open ended survey whose main purpose to get feedback about the courses from the customers to improve educational effectiveness. She provides a framework for structural coding of ULGQBD and brief road map for researchers who deals with ULGQBD.