The recent advancement of Web 2.0 enables people to exchange their opinions on a variety of topics. Among these discussions, the opinions of employees, customers, and investors are of great interest to companies. Insight into such perspectives can help managers make better decisions on business policies and strategy. However, assessing online opinions is a nontrivial task. The high volume of messages, casual writing style, and the significant amount of noise require the application of sophisticated text mining techniques to digest the data. Previous research has successfully applied sentiment analysis to assess online opinions on specific items and topics. In this research, we propose the integration of topic analysis with sentiment analysis methods to assess the public opinions expressed in forums with diverse topics of discussion. Using a Wal- Mart-related Web forum as an example, we found that combining the two types of analysis can provide us with improved ability to assess public opinions on a company. Through further analysis on one cluster of discussions, several abnormal traffic and sentiment patterns were identified related to Wal-Mart events. The case study validates the propose framework as an IT artifact to assess online public opinion on companies of interest. Our research promotes further efforts to combine topic and sentiment analysis techniques in online research supporting business decision making.