This study examines user sentiment toward the e-wallet applications DANA
and GoPay, which are widely used today. The dataset comprised 10,000 reviews
from Google Play Store and X for each platform between May 1 and June 1, 2025,
of which 6,457 DANA reviews and 7,274 GoPay reviews remained after data
cleaning. Sentiment classification using the IndoBERT model achieved average
accuracies of 93 % for DANA and 88 % for GoPay, confirming the robustness of
this approach. The analysis revealed a dominance of positive sentiment (37.4 % for
GoPay and 34.7 % for DANA), followed by neutral (31.0 % and 32.6 %) and
negative sentiment (31.6 % and 32.7 %), indicating substantial user satisfaction.
Evaluation of E-Service Quality dimensions showed strengths in reliability (49.1 %
for GoPay; 37.4 % for DANA) and responsiveness (38.0 % and 34.9 %), but
highlighted critical weaknesses in site organization for GoPay (58.8 % negative)
and in user friendliness (35.7 % negative) and personal needs (34.4 % negative)
for DANA. Topic modeling further identified positive themes—responsive
interfaces, transaction speed, promotions/cashback—as well as neutral-to-negative
themes related to technical support requests, transaction failures, security
concerns, and performance issues. These findings provide an empirical foundation
for DANA and GoPay to devise targeted strategies to enhance UX/UI, system
stability, and customer support, ultimately strengthening user satisfaction and
loyalty.
Keywords: Sentiment Analysis, DANA, GoPay, IndoBERT, Topic Modeling