Using data mining approach in new product development process: An application on smartphones

In today's world where information technologies are developing rapidly, the use of a systematic new product development process is very important in terms of the competition of enterprises. With the data science and artificial intelligence algorithms shown within the innovative technologies, collecting, organizing and analyzing data from various external sources such as social media, revealing the customer needs and putting this into the product development process in a systematic way can create great benefits for the enterprises. Therefore, within the scope of the study, it is aimed to present an integrated study that reveals customer demands from social media data and the difference between these demands, and also gives a development proposal for the selected product (smartphones) during the new product development process. For this purpose, the study was carried out in four stages. Firstly, Twitter data was collected with GetOldTweets application. Data Preprocessing and Topic Modeling was performed using LDA (Latent Dirichlet Allocation) method. Later, for Sentiment Analysis, training was carried out with Logistic Regression, Naive Bayes, Random Forest and XGBoost methods, and customer comments were scored with the Random Forest method, which gives the best accuracy. In the third stage, in order to obtain the Quality Function Deployment (QFD) method parameters used in the new product development process, the scores obtained as a result of the emotion analysis were reduced and Kano Transformation procedures were performed. Finally, the Quality House was obtained and suggestions for the smartphone were obtained. As a result of the study, eleven customer needs were determined by Topic Modeling, and their importance levels were obtained through Emotion Analysis and Kano Transformation. At the end of the application on smart phones, the three most important features in the development of these products were determined as Security, Technical Service Network, Operating System and Virtual Assistants. While the most important customer requirement was Quality Sense for the smartphone selected for development, with the effect of competitor parameters, the most important technical requirement was determined as the Application Processor. The study has shown that with the adaptation of innovative technologies to the new product development process, stronger, cheaper and faster inferences can be obtained.