Comparative Topic and Keyword-Based Sentiment Analysis from Indonesian YouTube Comments Between Thrifting and Local Clothing Brands – Bridging the Gap in Public Perception and Business Opportunities

VINCENTIA ROSELYND JEANNETE DHIAN RASTYANINGRUM, Ir. Yun Prihantina Mulyani, S.T., M.Sc., Ph.D., IPM., ASEAN Eng.

2024 | Skripsi | TEKNIK INDUSTRI

Pada awal tahun 2023, sebuah isu menarik perhatian publik di Indonesia setelah pemerintah memperkenalkan Peraturan Menteri Perdagangan No. 40 Tahun 2022 tentang Barang yang Dilarang Ekspor dan Impor. Toko barang bekas dan pasar loak menjadi tempat populer untuk menjual dan membeli barang bekas, yang mana khususnya menguntungkan individu dari kelas bawah dan menengah bawah. Namun, praktik thrifting di Indonesia yang semakin populer terutama di kalangan anak muda, memicu perdebatan mengenai dampaknya. Model bisnis thrifting yang bergantung pada pakaian bekas impor, menawarkan peluang untuk penjualan yang menguntungkan. Meskipun demikian, kekhawatiran muncul mengenai dampak impor pakaian bekas terhadap stabilitas industri tekstil lokal, dengan penelitian menunjukkan potensi kerugian PDB tahunan sebesar 11,87 miliar rupiah akibat meningkatnya popularitas thrifting yang berasal dari impor pakaian bekas. Akibatnya, pemerintah Indonesia menganggap aktivitas thrifting yang menjual pakaian impor sebagai ancaman bagi industri tekstil dan pakaian lokal.

Beberapa studi sebelumnya telah meneliti fenomena thrifting, perkembangan merek pakaian lokal, dan perbandingan keduanya menggunakan pengumpulan data terstruktur, yang cenderung menghasilkan kesimpulan yang mungkin mengabaikan aspek-aspek yang sedang dipelajari. Oleh karena itu, penelitian ini menerapkan pendekatan yang lebih eksploratif, yaitu text-mining, menggunakan Latent Dirichlet Allocation dan BERT Sentiment Analysis, dengan dataset yang diambil dari YouTube. Sebanyak 14.481 komentar dikumpulkan dari video YouTube tentang thrifting dan merek pakaian Indonesia.

Tujuh dan delapan topik diidentifikasi dari dataset thrifting dan lokal. Analisis sentimen dilakukan pada tingkat topik dan kata. Kedua analisis menyimpulkan sentimen yang lebih positif terhadap merek pakaian lokal. Dataset thrifting secara eksklusif membahas kesadaran, legalitas, bea cukai, dan kekhawatiran terhadap bisnis kecil, terutama aspek kebersihan. Sementara itu, kebanggaan lokal dan inovasi hanya ditemukan dalam dataset lokal. Meskipun ada perbedaan dalam kata kunci yang dihasilkan, analisis cakupan topik menggunakan adaptasi sembilan blok bangunan model bisnis dalam empat pilar mengungkapkan bahwa kedua dataset membahas topik terkait produk. Kata kunci yang tumpang tindih pada aspek produk dapat diekstraksi dari topik terkait. Ditemukan bahwa banyak kata kunci yang tumpang tindih menunjukkan preferensi pelanggan, perasaan, dan emosi terhadap produk pakaian. Perhitungan Indeks Peluang juga mengarah pada kesimpulan bahwa untuk bersaing dengan thrifting, bisnis pakaian lokal harus mempertahankan dan meningkatkan area produk, terutama terkait kualitas dan kesan pakaian. Bisnis lokal juga dapat memanfaatkan tingginya tingkat dukungan publik produk dalam negeri dengan penekanan pada promosi untuk memakai produk lokal.

In early 2023, an issue gained public attention in Indonesia following the government's introduction of Minister of Trade Regulation No. 40 of 2022 concerning Goods Forbidden for Export and Import. Thrift shops, garage sales, and flea markets emerged as popular spaces for selling and purchasing secondhand goods, particularly benefiting individuals from lower and lower-middle-class. However, thrifting practice in Indonesia has become increasingly popular, especially among the youth, and has sparked debates regarding its implications. The business model of thrifting, which mainly depends on imported used clothes, presents opportunities for profitable resale. Yet, concerns have been raised about the impact of importing secondhand clothes on the stability of the local textile industry, with research indicating potential annual GDP losses of 11.87 billion rupiah due to the growing prevalence of thrifting, particularly the import of used clothing. Consequently, the government of Indonesia perceives thrifting activities selling imported clothes as a threat to the local textile and clothing industries.

Several previous studies have studied the thrifting phenomenon, the development of local clothing brands, and the comparison of both adopted a structured data collection, which tends to result in a conclusion that may overlook aspects being studied. Therefore, this research applied a more explorative approach, namely text-mining, using Latent Dirichlet Allocation and BERT Sentiment Analysis, with datasets captured from YouTube. 14,481 comments were collected from YouTube videos about thrifting and Indonesian clothing brands.

Seven and eight topics were identified from thrifting and local datasets, respectively. Sentiment analysis was performed at the topic and keyword levels. Both analyses conclude a more positive sentiment towards local clothing brands. The thrifting dataset exclusively discusses awareness, legality, customs duty, and concerns around small businesses such as the cleanliness aspect. Meanwhile, local pride and innovation are only found in the local dataset. Although there are differences in the generated keywords, a topic coverage analysis using the adaptation of nine business model building blocks within four pillars reveals that both datasets discuss product-related topics. The overlapped keywords can be extracted from the related topics. It is found that many overlapped keywords signify customer preference, feelings, and emotions toward clothing products. Calculating the Opportunity Index also leads to the conclusion that to compete with thrifting, local clothing businesses should consider maintaining and improving the product area, especially regarding the quality and impression of the clothing pieces. Local businesses can also optimize the marketing strategy with accentuation in promoting to wear locals due to the high level of public approval in supporting local businesses.

Kata Kunci : Thrifting, Local Clothing Brand, Text Mining, LDA, Sentiment Analysis, BERT, Opportunity Index