Author: Dora Tse
Organizations that have been doing data analytics for a while may counter problems on the performance of dashboards and long waiting of query result, sometimes the slow performance affects other reports which share the same workspace in BI tool. Hence, other users cannot open reports and it will delay the queries until the resources is released. The root cause of it normally deal with gigantic dataset size and lots of on-the-fly calculations .
At this scenario, we normally would tune the report by scaling down the report size by splitting the dataset which was shared across reports, sometimes to avoid table aggregation. The biggest gain of our tuning once was to optimize way of calculations which we have tried end up reducing calculation time by 70% from one sales dashboards that used to contain 90 complex calculations. Besides, sometimes, performance lag related to dashboard display logic and we would resolve it by merging the various conditions in a measure as a true or false indicator for each record in the table to improve performance. This kind of data model redesign and query tuning we did would end up reducing query time by 50% and improve dataset size by another 50% easily which organization find our capabilities to tune fast the engine (BI analytics) would help strengthen their data foundation for more potential in future.
Best practices in BI design and data pipelines implementation are critical success factor for any BI Analytics project, each of the best practice we apply can significantly improve the speed of your final dashboard and report at end everytime . Faster and proper dashboards can speedup your decision through quicker data insights.
Special thanks to few project teams that are recently successful in performance tuning for our various clients, they are amazing!
Explora is not only your data partner, we are happy to be your optimization and tuning expert for your Analytics dashboards and data pipelines. If you have analytics performance concern, Explora will be your solution.
