ADVANCED COMPUTING APPLICATIONS IN BI DASHBOARDS: IMPROVING REAL-TIME DECISION SUPPORT FOR GLOBAL ENTERPRISES

This systematic review synthesizes how advanced computing architectures improve real time decision support in business intelligence dashboards for global enterprises. Following PRISMA, we screened major scholarly databases and citation networks, applied predefined eligibility criteria, and extracted methodological and performance data across pipeline stages from ingest to visualization. The final corpus comprised 115 peer reviewed studies. The evidence converges on a portfolio approach rather than a single technology: event time streaming with watermarks and stateful windows consistently lowers tail latency and staleness; deterministic, log centric materialization stabilizes results under late arrivals; hybrid transactional analytical processing reduces stale reads and compresses refresh windows; GPU accelerated SQL and fused operators lift interactive aggregation performance; and edge or fog placement trims “as of” lag where WAN variance is high. Cloud native orchestration and serverless patterns add elasticity and cost control for bursty workloads when scaling signals reflect workload semantics. Equally, governed semantic layers, knowledge graphs, lineage, and constraint validation reduce metric drift and reconciliation time, which raises sustained dashboard adoption. Privacy preserving telemetry and local anonymization enable cross border analytics with modest overhead, while stronger cryptography is reserved for narrow aggregate use cases. We provide a taxonomy that maps paradigms to capabilities, an evidence map linking mechanisms to outcomes, and pattern playbooks with practical SLO targets for P95 latency, freshness, and reliability. Limitations include workload heterogeneity and optimism in vendor authored cases, which we address through sensitivity analyses. Overall, assembling complementary paradigms with explicit semantics and governance yields durable, decision relevant gains for global BI dashboards.