Localization for embedded analytics is coming to Holistics. You build a dashboard once, add translations for each language, and each embedded viewer sees it in their own language.
If your customers speak different languages, they expect their dashboards in their language. Today, the only way to do that is to build and maintain a copy of each dashboard per language, which breaks down as you add languages.
System texts: buttons, menus, and error messages are translated by Holistics.
Dashboard texts: titles, text blocks, and field labels are translated from translation files you keep in your AML project (settings/locales/fr-FR.json).
Data values:H.current_locale lets you pick the right translated column from your database in your AML models.
One setting per viewer: pass settings.locale in the embed payload, and Holistics applies the right translations for that viewer.
AI chat notifications are now available in Holistics AI. You get notified when a conversation needs your attention, so you can work on something else while long tasks run.
Some AI tasks take a few minutes to run. Switch to another tab and you won't know when AI is done, or that it has been waiting on your answer the whole time.
You can now choose the color of the "Others" group in your charts. The "Others" group often holds more data than any single series, but until now you couldn't change its color. Mute it so your main series stand out, or make it bold when the long tail is the story.
Admins can now customize Home for everyone in their organization. Show your most important dashboard in full view right on Home, pin others as buttons, and choose which sections appear, in the order you want.
Many teams build a "start here" dashboard, like a KPI overview, a report glossary, or an onboarding page. But it sits one click away from Home, so occasional users rarely find it. Custom homepage puts it where everyone lands.
Pinned dashboards, in full view or as buttons: Show one Canvas dashboard in full view, so users can filter, drill, and explore it without leaving Home. Pin any others as buttons that open them in one click. Dashboards you've already pinned stay as buttons.
Section control: Show, hide, and reorder Home's sections: Pinned dashboards, Recents and favorites, and Browse.
Permissions apply: Everyone sees pinned dashboards with their own permissions and row-level access. Users who don't have access to a dashboard won't see it on Home.
Dashboard viewers can now hover over a column header in a table or pivot table and see a viewer-facing description of the field. This includes viewers on shareable and embed links.
Previously, field descriptions were only visible in Explore's field selector. Business users rarely opened it, and users on shareable or embed links couldn't open it at all, so the context analysts wrote never reached the people reading the dashboard.
Showing semantic-layer descriptions as-is isn't always safe either. The same description is often written for several audiences at once (AI, other analysts, business users), and can contain technical details or internal notes that aren't meant for clients.
To solve this, we've introduced the viewer-facing description: a description on each field in a table or pivot table, written for the people reading your dashboards and kept separate from the semantic-layer description:
Semantic-layer description: written in your models, visible only to users with dataset access, same as today.
Viewer-facing description: what dashboard viewers see in the tooltip. Always visible to anyone who can view the dashboard.
You decide where each viewer-facing description gets its content from:
Inherit from the semantic layer: if your semantic-layer descriptions are already written for business users and are safe to show to external viewers, turn on "Inherit field descriptions" in Project Settings. Every viewer-facing description then inherits its content from the field's semantic-layer description, so you don't have to write them again for viewers. You can also make exceptions for specific fields:
Force inherit: if only a few descriptions are ready for viewers, keep the project setting off and turn on inheritance for just those fields.
Don't inherit: if a few descriptions contain internal notes or technical details, keep the project setting on and turn off inheritance for just those fields.
Write a custom description: when a field shouldn't inherit, write business-friendly text for it on the visualization instead. The semantic-layer description stays untouched.
Default setting: for existing projects, "Inherit field descriptions" is off, so your dashboards look the same until you turn it on or add custom descriptions. New projects start with it on.
Configuration: the project setting is available in Project Settings. Field-level overrides and custom descriptions are configured in AML.
Empty descriptions: if a field has no inherited or custom content, no tooltip is shown.
Data Lineage shows what a model, dataset, or dashboard is built from and what depends on it, in a panel below the code in the IDE. Because it is compiled from your own project code, it stays accurate, so you can change things with confidence and trace any number back to its source.
You can now convert between Apache Ossie and AML in both directions. Apache Ossie (formerly Open Semantic Interchange, or OSI) is the open, vendor-neutral format for exchanging semantic models between tools.
Previously, a Holistics semantic layer lived only in AML, with no standard way to move it to or from another tool. Now you can:
Export an AML dataset to an Ossie model, so any tool that reads Ossie can use your Holistics definitions.
Import an Ossie model as AML files, so a model built in another tool comes into Holistics without being rewritten.
The converter is open source and lives in the Apache Ossie converters repository. It runs on your machine with the Holistics CLI, and needs no Holistics account or warehouse connection. Holistics is also listed on the Apache Ossie ecosystem page.