PENERAPAN PYTHON UNTUK MENGIDENTIFIKASI POLA KONSUMSI PRODUK FASHION MELALUI VISUALISASI DATA

SALSABILLA, CLAUDIA MAHARANI (2025) PENERAPAN PYTHON UNTUK MENGIDENTIFIKASI POLA KONSUMSI PRODUK FASHION MELALUI VISUALISASI DATA. Diploma thesis, UIN Raden Intan Lampung.

	PDF Download (3MB)
	PDF Restricted to Repository staff only Download (3MB)

Abstract

ABSTRAK Pertumbuhan pesat industri Fashion digital di Indonesia mendorong kebutuhan akan pemahaman yang lebih mendalam terhadap pola konsumsi konsumen, namun analisis perilaku konsumen berbasis data transaksi e-commerce masih terbatas dalam praktik bisnis saat ini. Penelitian ini bertujuan untuk mengidentifikasi pola konsumsi produk Fashion melalui analisis data dengan pendekatan clustering dan visualisasi data interaksi. Data diperoleh 211 responden pengguna platform e-commerce melalui survei kuesioner. Penelitian ini menggunakan metode CRISP-DM (Cross-Industry Standard Process for Data Mining) dengan teknik segmentasi menggunakan algoritma DBSCAN dan reduksi dimensi PCA berhasil mengelompokkan konsumen ke dalam tiga segmen utama berupa Low Engagement (frekuensi belanja rendah), Discount Lovers (sangat responsif terhadap diskon dan rating), dan High Frequrncy (frekuensi pembelian rutin bulanan). Validasi hasil menunjukkan kualitas segmentasi yang kuat melalui nilai Silhouette Score sebesar 0.540 dan Indeks Calinski-Harabasz sebesar 127,6. Hasil analisis berhasil mengidentifikasi empat segmen konsumen utama: Low Engagement (45,9%), Discount Lovers (29,4%), High Frequency (11,4%), dan High Value (3,8%). Hasil segmentasi ini kemudian divisualisasikan dalam dashboard interaktif menggunakan tableau untuk mendukung keputusan strategis. Penelitian ini memberikan kontribusi praktis berupa rekomendasi strategi pemasaran yang terdiferensiasi untuk setiap segmen, sehingga dapat meningkatkan efektivitas pemasaran dalam industri Fashion digital. Kata Kunci : Clustering, DBSCAN, Python, E-commerce Fashion, Segmentasi Pelanggan. ABSTRACT The rapid growth of the digital Fashion industry in Indonesia has increased the need for deeper insights into consumer behavior. However, the application of data-driven analysis to understand e commerce consumption patterns remains limited in practice. This study aims to identify Fashion product consumption patterns through the implementation of data analysis utilizing clustering techniques and interactive data visualization. Data were collected from 211 respondents who are users of e-commerce platforms in Indoensia theough a questionnaire-based survei This research uses the method CRISP-DM (Cross-Industry Standard Process for Data Mining) framework. The segmentation process employed the DBSCAN clustering algorithm and dimensionality reduction with PCA, resulting in three main cunsumer segements Low Engagement (low purchase frequency), Discount Lovers (highly responsive to discounts with ratting), and High frequency (regular monthly purchasers). The results were validated using a Silhouette Score of 0.540 and a Callinski-Harabasz Indeks of 127,6. The analysis successfully identified four main consumer segments: Low Engagement (46.9%), Discount Lovers (29.4%), High Frequency (11.4%), and High Value (3.8%). These segmentation result were then visualized in an interactive dashboard using tableau to support startegic decisionn making. This research provides practical contributions in the form of differentiated marketing startegy recommendations for each segment, which can enhance marketing effectiveness in the digital Fashion industry. Keyword : Clustering, DBSCAN, Python, E-commerce Fashion, Customer Segmentation.

Actions (login required)

View Item