Suggestions for Improving Dashboard Performance with Big Datasets

[JesicaW]

Hello Everyone,

I’m currently working on a dashboard that involves processing and visualizing very large datasets (around 50 million rows), and I’m encountering some performance issues. The dashboard loads slowly, and some widgets take a long time to render or fail to load altogether. I’m hoping to get advice on best practices for optimizing both the dashboard’s performance and the speed of querying these large datasets from the community.

Here’s a bit more context on my current setup:

• I’m using Sisense version [Insert version here].
• The data is sourced from [Insert your data source, e.g., SQL Server, Redshift, Google BigQuery].
• I’ve implemented aggregations to reduce data size but still see lag when filtering.
• My widgets are a mix of bar charts, line graphs, and pivot tables, each pulling different slices of the dataset.
• I’ve tried adding filters to the dashboard to limit the data being loaded, but the performance improvement is minimal.

Here are a few specific questions I have:

1. What are the most effective methods for improving the performance of dashboards that rely on large datasets in Sisense?
2. Are there any strategies or configurations I should consider at the data model level that might help streamline query execution?
3. Could using ElastiCubes versus live data models influence performance in a significant way for datasets of this size? If so, how should I decide which approach is best?
4. Any recommendations for managing slow-loading widgets? Would simplifying the widget calculations help? 

I’d appreciate any insights or resources you could share that might help me tackle these gen ai challenges. Thank you in advance.

[DRay]

Hello @JesicaW,

Thank you for reaching out. There is a lot of content out there for optimizing widget performance. What resources have you already looked at?