This study conducted unstructured big data analysis to confirm consumers' perception of cosmetics in the face of increasing non-face-toface consumption due to COVID-19 pandemic. To this end, keywords related to cosmetics were collected through portal sites such as Naver, Daum, and Google using the social network matrix program called Textom. The collection period was set from Jan 2020 to DEC 2021, and a total of 18,672 keywords were collected, and a total of 60 keywords were used for the study by refining unnecessary keywords. The results are as follows. First, performing frequency, TF-IDF analysis, important keywords such as skin, basic makeup, recommendation, use, and brand were presented. Next, semantic network analysis showed that degree centrality was skin, basic makeup, recommendation, use, and brand, closeness centrality was CHANEL, authentication, export, and customize, betweeness centrality was skin, use, brand, products, and basic makeup. Finally, CONCOR analysis resulted in a cluster of five groups: cosmetic attributes, ingredients, products, sales, and using cosmetics. These analysis results confirm consumer perceptions related to cosmetics, key components and sales channels for cosmetics In addition, it is judged that it will propose meaningful implications for establishing effective data presentation and marketing strategies for research related to demand for cosmetics after COVID-19 pandemic.