Combining content analysis and neural networks to analyze discussion topics in online comments about organic food

Consumers increasingly share their opinions about products in social media.However, the analysis of this user-generated content is limited either to small,in-depth qualitative analyses or to larger but often more superficial analysesbased on word frequencies. Using the example of online comments aboutorganic food, we suggest a three-step methodological approach of how latestdeep neural networks can scale up the insights of qualitative analyses. First, aqualitative content analysis defines a class system of opinions. Second, a pre-trained neural network, the Universal Sentence Encoder, uses this class systemto automatically classify the same data by finding similar opinions. Third, theautomatic classification results are evaluated based on several criteria. Wefind coherent results of qualitative and automated classification proving theability of Universal Sentence Encoder to classify text. After this validation,Universal Sentence Enconder can be used to classify larger data sets onorganic food. The suggested approach allows to scale up sample size whilemaintaining the detail of class systems provided by qualitative contentanalyses. The approach can be applied to different domains and supportconsumer and public opinion researchers as well as marketing practicionersin further uncovering the potential of insights from user-generated content.