Cube

Keep analytics discussions next to the insight they’re about. Cube now supports comments on Dashboards and individual charts. Teams can reply in threads, mention teammates to notify them in their inbox, react with emoji, and resolve threads when a question is settled. That keeps questions and decisions alongside the data that prompted them. How does your team keep analytics discussions connected to the work?

More ways to compare periods in Cube now covers filtered dates and multiple date fields in the same analysis. This demo shows how comparative analysis can be configured directly in Cube: - Compare a report that only has a date filter, even when the date is not shown as a breakdown column. - Let comparison presets follow the dashboard filter window users are actually viewing. - Compare one measure across multiple dates side by side, with clear column titles when the date context matters. It helps business users answer the next question faster: not just what happened, but how it changed. Which comparison period do your stakeholders ask for most often? #Cube #Analytics #Dashboards #BusinessIntelligence

Cube filter operators now cover calendar periods and empty text values. This demo shows how the new filters work across workbooks, dashboard filters, and published dashboards: - Filter time fields by month, quarter, or year without writing manual date logic. - Use empty and not-empty text filters for cleaner handling of missing values. - Move between point-and-click filters and Semantic SQL while keeping the controls editable. Where would calendar-period filters remove the most custom SQL from your workflow? #Cube #Analytics #SemanticLayer #BusinessIntelligence

Cube Semantic SQL now makes custom SQL filters easier to read and edit. This demo shows two related improvements in the query and filtering flow: - The Semantic SQL editor formats long queries into readable clauses instead of one horizontal line. - Custom SQL filters become easier to inspect when conditions span multiple lines. - Filter controls stay connected to the SQL, so teams can edit logic without turning it into an unreadable block. What SQL editing rough edge slows your analytics workflow the most? #Cube #SemanticSQL #AnalyticsEngineering #DataTools

Cube donut and pie charts give dashboards another compact way to show part-to-whole analysis. This demo walks through how to build and tune circular charts in Cube: - Create donut or pie visualizations from query results. - Compare category contributions at a glance instead of scanning a table. - Adjust the chart presentation so dashboard readers can quickly understand the split. Where do pie or donut charts still help your stakeholders make a faster decision? #Cube #DataVisualization #Analytics #Dashboards

An analytics agent should be able to build on the metric definitions your data team already maintains. With Snowflake CoCo and Cube, that workflow can stay intact: • Your team authors definitions in Snowflake Semantic Views. • Cube imports them into an executable semantic model. • CoCo discovers measures and dimensions, runs governed queries, and saves workbooks, reports, and dashboards through Cube. There are two connection paths. Direct MCP tools such as searchDataModel and runQuery keep CoCo in charge of the analysis while Cube handles query execution. The Cube CLI and open-source agent skills provide a more structured path for model exploration, query validation, and content creation. The important boundary is permissions. MCP calls run with the connected Cube user's access, so the agent works with the same governed model and policies as the people who will use its output. The result is not just an answer in a terminal. It is a dashboard a teammate can open, filter, change, and share. See the architecture and setup: https://lnkd.in/gDHiHPjX #AgenticAnalytics #Snowflake #SemanticLayer #DataEngineering

Cube small multiples make comparative analysis easier when one chart is not enough. This demo shows how to split a visualization into a grid of panels by dimension: - Compare the same trend across regions, channels, or other segments side by side. - Use a shared axis when direct comparison matters. - Switch to independent axes when each panel's shape needs more readable scale. What dimension would you split first when comparing performance across segments? #Cube #DataVisualization #Analytics #BusinessIntelligence

Cube's query and refresh controls are getting simpler. This demo shows UI improvements for running workbook queries and refreshing chart data: - Refresh charts from the freshness indicator dropdown in dashboards and workbooks. - Use clearer Run and Auto-run controls while building workbook queries. - Keep query execution explicit when you want control and automatic when iteration speed matters. Which query control do your users reach for most: manual run, auto-run, or refresh? #Cube #Analytics #Dashboards #ProductUpdate

Comments on Cube dashboards and charts keep feedback attached to the exact data teammates are discussing. This demo shows how to: • Start a thread on the whole dashboard or a specific chart, preserving the analytical context • Reply, react with emojis, and mention teammates so the right person can jump in from their inbox • Resolve a discussion without losing its history, or copy a direct link to share it elsewhere Where does dashboard feedback go on your team today?

The risky part of BI is not the Python model. It is the CSV export nobody can reproduce. Cube's new Python analysis feature points at a larger analytics pattern: teams increasingly need statistical methods inside the same governed workflow as dashboards and semantic-layer queries. SQL is still excellent for aggregations: cohorts, retention, period-over-period analysis, top-N lists, running totals. But forecasting, regression, significance testing, and clustering usually move somewhere else. A data scientist exports a CSV, writes a notebook, pastes a chart into a deck, and two weeks later the team has a reproducibility problem. The important idea here is not simply "Python in BI." It is that the script runs against the semantic SQL result from the workbook tab. The forecast or model output is tied to the same reviewed definitions that power the rest of the analytics surface. That matters as analytics becomes more agentic. Agents can ask for richer analysis, but richer analysis only helps when it stays connected to governed metrics, lineage, permissions, and saved artifacts other people can rerun. Read the post: https://lnkd.in/gXiefjab