Workbooks: From Purposeful Data to Interactive Decisions

Static dashboards and BI reports often create more questions than they answer. When a business user sees a number, their immediate next step is to ask "why?"—a question that sends analysts back to their tools, creating a slow, frustrating loop between data and decisions.

Arkham Workbooks are engineered to break this cycle. They are live, interactive dashboards built on a foundation of governed data, allowing users to not only see what's happening but to explore, filter, and ask follow-up questions in real time. By connecting directly to the curated Metric Store and live ML model outputs, Workbooks transform static reports into dynamic decision-making tools.

How They Work: From Governed Data to Visualization

The key to building a reliable dashboard is ensuring the data comes from a trusted, well-defined source. Workbooks enforce this by design, providing a low-code canvas for visualizing outputs from either the Metric Store or the ML Hub.

The workflow follows one of two paths:

  • For Business Intelligence: The journey begins in the Metric Store, where you define reusable business logic (e.g., SUM(revenue)). In a Workbook, you connect a panel to one of these pre-defined Metrics to visualize it.
  • For Model Analytics: The journey begins in the ML Hub, where you train and deploy a model. The model's predictions (output dataset) are now available as a source in Workbooks. You can build dashboards to monitor the model's performance over time or create "what-if" scenarios by changing input parameters.

🤖 AI-Assisted Analytics with TARS

TARS acts as an analytical co-pilot directly within the Workbook interface, empowering both builders and consumers to interact with data more intuitively.

For Builders: Accelerate dashboard creation by using natural language. > "TARS, create a new KPI panel showing total active_users for the last 30 days."

For Consumers: Go beyond the dashboard's static filters. Ask follow-up questions to dig deeper into the data. > "TARS, what was the peak revenue day last month?"

Key Features for Builders

  • Low-Code Interface: Build sophisticated dashboards without writing front-end code. The UI handles panel arrangement, chart configuration, and interactivity.
  • FuGoverned Data Sources: By building exclusively on the Metric Store or deployed model outputs, Workbooks ensure that all dashboards are governed by a single source of truth. This eliminates inconsistency and ensures stakeholders see the same, accurate numbers everywhere.
  • Reusable Business Logic: Build on top of the centralized Metrics you've already curated. This prevents logic from being duplicated and ensures consistency across all BI-related analytics.
  • Direct Model Integration: Seamlessly visualize the outputs of your deployed ML models. This is critical for monitoring model drift, evaluating accuracy, and explaining results to stakeholders.
  • Performance Optimization: Our platform automatically handles query caching to provide a fast user experience, with the option to bypass the cache when real-time data is critical.

Core Components

  • Panels: The building blocks of a workbook. Choose from various types including Markdown, KPIs, tables, and a wide range of charts (line, bar, donut, etc.).
  • Filters: Control the data displayed in panels. Global filters affect all panels, while local filters apply to a single panel.
  • Edit Mode: The builder-focused interface for configuring workbooks. Change layouts, modify panel types, rename axes, set colors, and define number formatting.
  • AI Platform Overview: Understand how Workbooks fit into the AI Platform ecosystem.
  • Metric Store: Define the governed, reusable business logic that powers your BI dashboards.
  • ML Hub: Create the deployable models whose outputs can be visualized and monitored in Workbooks.
  • TARS: Accelerate dashboard creation and analysis with a conversational AI co-pilot.