The availability of social media-based data creates opportunities to obtain information about consumers, trends, companies and technologies using text mining techniques. However, the quality of the data is a significant concern for social media-based analyses. The aim of this study was to mine tweets (microblogs) to explore trends and retrieve ideas for various purposes such as product development, technology and sustainability-oriented considerations. The core methodological approach was to create a classification model to identify tweets that contained an idea. This classification model was used as a pre-processing step so the query results obtained from the application programming interface were cleared from the messages that contained the search terms used in the query but did not contain an idea. The results of this study demonstrate that our method based on text mining, and supervised or semi-supervised classification methods, can extract ideas from social media. The social media data mining process illustrated in our study can be utilised as a decision-making tool to detect innovative ideas or solutions about a product or service and summarise them into meaningful clusters. We believe that our findings are significant for the sustainability, tech mining and innovation management communities.

基于社交媒体的数据的可用性为使用文本挖掘技术获取有关消费者、趋势、公司和技术的信息创造了机会。然而，数据的质量是基于社交媒体的分析的一个重要问题。本研究的目的是挖掘推文（微博）以探索趋势并检索出于各种目的（例如产品开发、技术和面向可持续性的考虑）的想法。核心方法论方法是创建一个分类模型来识别包含想法的推文。该分类模型用作预处理步骤，因此将从应用程序编程接口获得的查询结果从包含查询中使用的搜索词但不包含想法的消息中清除。这项研究的结果表明，我们基于文本挖掘和监督或半监督分类方法的方法可以从社交媒体中提取想法。我们研究中说明的社交媒体数据挖掘过程可用作决策工具，以检测有关产品或服务的创新想法或解决方案，并将其总结为有意义的集群。我们相信我们的发现对可持续发展、技术挖掘和创新管理社区具有重要意义。