Fiyas Mahananing Puri, Surya Tri Atmaja Ramadhani, Dewi Anisa Istiqomah, Vikky Aprelia Windarni, Kardilah Rohmat Hidayat

Analisis Rekomendasi Pembuatan Produk Menggunakan RStudio Dan Twitter (Studi Kasus : Git Solution)

Introduction

Analisis rekomendasi pembuatan produk menggunakan rstudio dan twitter (studi kasus : git solution). Tingkatkan rekomendasi produk e-commerce dengan analisis big data Twitter via RStudio. Studi kasus Git Solution, gunakan K-Means & SAW untuk rekomendasi personal. Hasil DB: 0,10%.

Abstract

Dalam lanskap korporat kontemporer, internet telah muncul sebagai alat fundamental untuk meningkatkan penjualan dan layanan pelanggan, terutama melalui penerapan sistem rekomendasi yang banyak digunakan khususnya dalam e-commerce. Namun, banyak perusahaan, termasuk PT GIT Solution, masih belum sepenuhnya memanfaatkan potensi big data dalam analisis pasar, dan masih bergantung pada data historis serta metode konvensional seperti follow-up atau kunjungan langsung. Hal ini menyebabkan kurang optimalnya pemanfaatan data untuk menghasilkan rekomendasi yang relevan dan personal bagi pelanggan. Penelitian ini menggunakan pendekatan analisis data Twitter dengan algoritma pengelompokan K-Means untuk mengelompokkan data dan metode perangkingan Simple Additive Weighting (SAW) untuk memberikan rekomendasi produk yang disesuaikan dengan preferensi pelanggan. Data yang digunakan diperoleh melalui crawling data Twitter, dan dianalisis menggunakan RStudio untuk mengidentifikasi tren serta preferensi konsumen. Hasil penelitian menunjukkan bahwa metode yang digunakan memberikan hasil evaluasi yang memuaskan dengan Indeks Davies-Bouldin (DBI) sebesar 0,10%, yang menunjukkan kualitas pengelompokan yang baik. Implikasi dari hasil ini adalah perusahaan dapat memanfaatkan analisis big data dari media sosial untuk meningkatkan kualitas rekomendasi produk, yang pada akhirnya dapat meningkatkan kepuasan pelanggan dan efektivitas pemasaran.

Review

This paper presents a timely and relevant study on leveraging big data from social media for product recommendations, specifically addressing the challenge faced by companies like PT GIT Solution that rely on conventional methods. The abstract clearly articulates the problem of underutilized big data in market analysis and proposes a solution using Twitter data. The chosen approach, combining K-Means for clustering and Simple Additive Weighting (SAW) for ranking, implemented in RStudio, is a pragmatic method for generating personalized product recommendations. The study's objective to move beyond historical data and direct follow-ups is well-justified in the current digital landscape. A significant strength of this research lies in its clear methodology and the practical application of established data mining techniques to a real-world problem. The use of Twitter data, a vast and dynamic source of public opinion, for identifying consumer trends and preferences is commendable. The reported Davies-Bouldin Index (DBI) of 0.10% provides a quantitative measure of the clustering quality, indicating a well-separated and compact grouping of data, which is crucial for effective recommendation systems. This result underpins the potential for companies to significantly enhance their market analysis and recommendation accuracy by integrating social media insights, thereby improving customer satisfaction and marketing efficacy. While the abstract presents a robust approach and promising results, a more comprehensive review would ideally delve into several areas for potential enhancement or deeper discussion. It would be valuable to understand the specifics of how "preference" was extracted and operationalized from raw Twitter data, including any pre-processing steps like sentiment analysis or topic modeling that might precede clustering. Furthermore, while K-Means and SAW are effective, a comparative analysis with other clustering algorithms (e.g., DBSCAN for varying density clusters) or ranking methods could provide insights into the generalizability and robustness of the proposed system. Future work could also explore the real-time implications of such a system, how it handles data noise inherent in social media, and whether its findings are generalizable beyond the specific case study to different industries or broader social media platforms. Overall, this paper offers a valuable contribution to the field of big data analytics for marketing.

Full Text

Comments

You need to be logged in to post a comment.