Online worlds offer a massive display of people's lives including their creative activities. In the present study, we investigated do-it-yourself (DIY) videos on YouTube to explore the type and prevalence of everyday creative behaviors. DIY is a term associated with the production of original and effective products, and, like YouTube, typically features amateur-generated content. As such, this setting allowed us to gain insight into the creative behavior that happens in our homes, garages, and front yards. Using two topic modeling approaches, structural topic model (STM) and BERTopic, we analyzed about 13,000 videos from three mainly English-speaking countries (Australia, the United Kingdom, and the United States) that were published between 2019 and 2021. In a final set of 5,618 videos including transcripts, STM identified 21 topics, such as Decoration, Garden, Interior Design, and Textiles, which are partly similar to, yet also extend established creativity domains. BERTopic identified 35 slightly more narrow topics of creative activities (e.g., Facemasks, Accessories, Candles) that offer additional insights into the actual diversity of everyday creative behavior. Further analyses uncovered plausible time dynamics of specific topics and examined the overlap and differences of the two chosen modeling approaches. Our results indicate that leveraging data-driven approaches to study creativity in online environments complements traditional methods in offering new insights into both the diversity and prevalent domains of everyday creative behavior.