Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal International Journal of Science and Society (IJSOC)
Della Anissa Putri Widodo
Universitas Singaperbangsa Karawang, Indonesia
Materials Science & Nanotechnology Social Sciences

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Optimizing Deepseek Chat API Costs Based on Log Pattern Analysis using Business Intelligence and CRISP-DM Della Anissa Putri Widodo; Nono Heryana; Azhari Ali Ridha
International Journal of Science and Society Vol 8 No 4 (2026): International Journal of Science and Society (IJSOC)
Publisher : GoAcademica Research & Publishing

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54783/ijsoc.v8i4.1764

Abstract

This study aims to analyze token usage and cost patterns of the DeepSeek Chat API and formulate cost optimization strategies based on historical log data. The study employs a Business Intelligence approach with the CRISP-DM framework through descriptive and diagnostic analysis of usage data from July 9 to November 30, 2025. The analysis covers token usage, cost, cost per token, usage patterns over time, correlation, and context caching. The results show a total usage of 81,612,706 tokens with a total cost of USD 11.212044 and an average daily usage of 562,846.25 tokens. The aggregate cost per million tokens was USD 0.137381, while the Pearson correlation between token usage and cost was 0.791494. Context caching analysis showed a cache hit rate of 83.97% and a cache miss rate of 16.03%. Based on the analysis, usage patterns with relatively higher costs can be identified to support token volume monitoring and the utilization of context caching as part of cost optimization strategies. The results are presented through a Business Intelligence dashboard as a tool for monitoring and evaluating DeepSeek Chat API usage.
Co-Authors Azhari Ali Ridha Heryana, Nono