Given the increasingly diverse preferences of customers, utilizing the qualitative information obtained from comments written in the entry fields of questionnaires and those posted on websites is becoming more important. Understanding the thoughts and feelings of a targeted group of people can lead to providing better services and making more accurate decisions. In order to analyze this large amount of data, however, the person in charge must review the comments one by one and understand the intended meaning, a process that requires a considerable amount of time. In addition, if the data is analyzed by multiple people, the results may vary according to each person's experience and knowledge. This is why visualizing the data quantitatively and comprehensibly is important.
To address this challenge, Fuji Xerox is conducting research on natural language processing technology that can use computer analysis in addition to human interpretation to effectively collect and analyze the text data written in documents. Using this technology, a massive amount of text data can be automatically categorized and organized according to the flow described in Fig. 1 so that the data can be visualized quantitatively as well as comprehensibly. For example, visual graphs such as topic graphs (Fig. 2) and mosaic plots (Fig. 3) can be used to visualize data. This allows the analyst to analyze data without having to interpret any of the text data, which allows for efficient understanding of the overall trends and changes in the data for better evaluation.
We can see that words "Power," "OK," "Voice," "Manner," and "Disturbing" are related to the word "Train" in this text because these topics are connected by pink arrows (a). "Train" and "Manner" are connected by arrows in both directions (a, b), thereby indicating that both topics are closely related.
Below is an example of the results of an analysis using text data from questionnaires on where people buy foundation (Fig. 4). The text data written in the entry fields is automatically categorized and organized using computer analysis. This data is shown quantitatively using a topic graph and a mosaic plot.
This technology can be used to analyze non-quantitative questionnaires and customer feedback, meaning that the results can be used for many other challenging applications such as improving the quality of service in restaurants and hotels or faster product planning.