Start Date
12-13-2015
Description
Today, people are more and more active in social networks and communicate via text massages, images or “likes”. Especially images are used to assist a person to provide their opinion. Images show the daily life or things that interest people (e.g. van House 2011). Thereby a huge amount of information is provided. The evaluation of images would enable a comprehensive classification of the consumer. Therefore the technologies of image classification like support vector machines (SVM) are needed. This study provides an approach to analyze images for market research. For this, we conducted a holiday survey. We asked 433 people about typical holiday activities and to upload their favorite holiday images. Overall 1,348 images have been uploaded. With the help of SVM, we could classify the images and evaluate particularly useful features. The study’s findings advance the possibilities of market research methods and provide numerous implications for researchers and practitioners.
Recommended Citation
Daniel, Ines and Baier, Daniel, "Towards Lifestyle Segmentation via Uploaded Images from Surveys and Social Networks" (2015). ICIS 2015 Proceedings. 2.
https://aisel.aisnet.org/icis2015/proceedings/SocialMedia/2
Towards Lifestyle Segmentation via Uploaded Images from Surveys and Social Networks
Today, people are more and more active in social networks and communicate via text massages, images or “likes”. Especially images are used to assist a person to provide their opinion. Images show the daily life or things that interest people (e.g. van House 2011). Thereby a huge amount of information is provided. The evaluation of images would enable a comprehensive classification of the consumer. Therefore the technologies of image classification like support vector machines (SVM) are needed. This study provides an approach to analyze images for market research. For this, we conducted a holiday survey. We asked 433 people about typical holiday activities and to upload their favorite holiday images. Overall 1,348 images have been uploaded. With the help of SVM, we could classify the images and evaluate particularly useful features. The study’s findings advance the possibilities of market research methods and provide numerous implications for researchers and practitioners.