Abstract
We analyze a tweet system in a software development company. This tweet system has been used for many years as a informal communication tool by people at main office, at local office and at customers. We analyze the term that appears in this tweet system, and extract the topic. We think that we can search for the symptom of the quality of overtime hours by searching for the relation between the change of topic and overtime hours. By correlative analysis of the difference from the stable state of the transition of the topic and overtime hours, we try to understand the symptom whether they work hard positively or they need a rest. In addition, we try to make figure the transition in the topic.
