<![CDATA[Abstrak - Industri telekomunikasi menghadapi tantangan besar dalam mempertahankan pelanggan karena meningkatnya persaingan. Churn pelanggan, yaitu perpindahan ke kompetitor, menimbulkan kerugian finansial yang signifikan. Penelitian ini mengembangkan aplikasi prediksi churn berbasis web menggunakan algoritma Random Forest dan framework Streamlit. Data yang digunakan berasal dari 7043 pelanggan dengan 21 atribut. Proses penelitian mencakup pengumpulan data, data cleaning, label encoding, pembagian data (80:20), pelatihan model, serta evaluasi. Hasil menunjukkan bahwa model Random Forest mencapai akurasi 78.82%, dengan precision 83% dan recall 90% untuk kelas non-churn, serta precision 63% dan recall 48% untuk kelas churn. Fitur yang paling berpengaruh terhadap churn adalah Contract type, Tenure, TotalCharges, dan MonthlyCharges. Aplikasi ini menyediakan antarmuka interaktif yang memudahkan pengguna dalam menganalisis data dan memprediksi churn, membantu perusahaan telekomunikasi mengidentifikasi pelanggan berisiko tinggi dan meningkatkan strategi retensi.Kata kunci : Churn Pelanggan; Random Forest; Telekomunikasi; Streamlit; Abstract - The telecommunications industry faces major challenges in retaining customers due to increasing competition. Customer churn, or the shift of customers to competitors, causes significant financial losses for companies. This study developed a web-based customer churn prediction application using the Random Forest algorithm and the Streamlit framework. The dataset used consists of 7,043 customers with 21 attributes. The research process includes data collection, data cleaning, label encoding, data splitting (80:20), model training, and evaluation. The results show that the Random Forest model achieved an accuracy of 78.82%, with precision of 83% and recall of 90% for the non-churn class, as well as precision of 63% and recall of 48% for the churn class. The most influential features in predicting churn are Contract type, Tenure, TotalCharges, and MonthlyCharges. The developed application provides an interactive interface that allows non-technical users to analyze data and predict churn, helping telecommunications companies identify high-risk customers and improve retention strategies.Keywords: Customer Churn; Random Forest; Telecommunications; Streamlite;]]>
