Nine enterprise data / AI offers — Amazon Redshift, Microsoft Fabric, Google BigQuery, Snowflake, Databricks, Dataiku, Oracle Autonomous AI Database, IBM watsonx.data and Cloudera CDP — benchmarked, correlated, compared, heat-mapped, dependency-mapped and gap-analysed from your comparison report, with a global cloud data-center footprint.
There is no single winner across warehouse, lakehouse, operational databases, data science and cloud control — the shortlist should follow workload and control requirements.
Re-weight the eight capability dimensions to match your priorities and watch the ranking, the radar profile and the capability bubbles respond.
Weighted share of the maximum possible score (feature breadth).
Toggle platforms to overlay their eight-dimension profiles.
Every bubble is one platform × dimension score; bubble area grows with the score. Regroup the packing to see strengths by platform or market depth by dimension.
Head-to-head duels, the cross-platform capability matrix from the report, the nested treemap of all scores and the share of total.
Pick any two platforms. Bars extend from the centre.
Dimension family → dimension → platform; tile area = score.
Native data engines versus broader cloud portfolio capability (from the report). Filter by any keyword.
Switch the lens: share of total benchmark score, a platform's dimension mix, or data-center regions by vendor.
Do platforms that are strong in one dimension tend to be strong in another? Pearson correlation across the nine platforms, a configurable scatter, and chord diagrams of co-strength and platform overlap.
Pearson r across 9 platforms (n=9 — directional, not statistically conclusive).
Choose two dimensions. Overlapping platforms are fanned out for readability.
Ribbon width = number of platforms scoring ≥ 4 in both dimensions.
Ribbon width = Σ of the lower score per dimension (shared capability depth).
Where each platform is strong, where it falls short of the ideal, and where it leans hardest on external or customer-owned layers.
Platforms × dimensions (1 = limited · 5 = broad, mature native capability). Right: platform total / 40.
5 − score. Hotter = larger gap to a broad, mature native capability.
How much each platform relies on separate products or customer-owned controls per layer (1 low – 3 high). Qualitative, derived from the report's dependency notes.
A production data/AI platform is a graph of managed services and customer-owned controls. Arrows indicate a dependency, not a data-flow guarantee.
Platforms (top-left arc) link to the layers and services they depend on. Hover any node to isolate its connections.
Each platform fans out to its services, its headline strength (green) and its headline gap (red). Drag nodes; they spring back into shape.
The largest gaps are architectural rather than missing checkboxes. Treat each as a validation item in the selection process.
Stacked 5 − score, per dimension. Shorter bar = closer to broad native capability.
Average gap by dimension across all nine platforms.
Follow a primary requirement to its shortlist, then to the gap themes each shortlisted platform most needs validated.
Requirement → shortlist mapping is from the report's selection guidance. Platform → gap mapping is derived from the report's gap-analysis text.
Primary requirement → shortlisted platform → the decision test to run. Hover to read the path.
A Voronoi partition of the world: each cell belongs to the nearest cloud region — a proxy for how close each vendor's capacity sits to users. Hover for the nearest region; scroll to zoom, drag to pan.
Toggle vendors to recompute the partition.
Added context — not in the source report. Approximate public cloud region locations (city level) compiled from vendor region lists; illustrative and not an exhaustive or verified facility inventory. Snowflake, Databricks and Dataiku run on these hyperscalers; Cloudera and IBM also support customer-managed / hybrid footprints. Verify region and SKU availability at procurement time.
Count of mapped regions.
Share of each vendor's mapped regions by macro-region.
Key milestones per platform across three eras. Scroll to zoom, drag to pan.
One lane per platform; stems mark milestones.
Added context — not in the source report. Approximate public-record dates for orientation; era labels are an editorial framing. Verify before external use.