This paper aims to analyze trends of social, cultural change by observing target words and their co-words extracted from a large-scale newspaper corpus, Trends 21. The Trends 21 corpus is a collection of four major Korean daily and national newspapers(Chosun, Dong-a, Joongang, and Hankyoreh) issued from the year 2000 and it has developed over a long period of time. In order to social change, we propose that a statistical method such as a t-score should be introduced for selecting target words and extracting their co-words appearing in the same paragraph. In addition, we construct network to show the relations between target words and their co-words explicitly. Also we interpret that degree of increase/decrease of target word and their co-words reflect social change. Finally we propose that some trends of ``family, friend`` can be captured based on these framework.
