Visual Object

Visual Object Overview#

The Reasoning Flows layer for visualization, reporting, and delivery of data insights.

Purpose#

The Visual Object type in Reasoning Flows enables the creation of visual representations, reports, and interactive applications based on processed and transformed data. These objects help teams illustrate trends, patterns, and insights through notebooks, custom applications, and automated delivery systems — turning data into shareable intelligence.

Note: Some objects available under Visual Object also appear under other object types (such as Transform & Prepare or AI & Machine Learning). The distinction is based on intent: when your goal is visualization, reporting, or insight delivery, use Visual Object.

Where It Fits in Reasoning Flows#

In the Reasoning Flows architecture:

1. Extract & Load brings data into the platform.
2. Repository Tables register datasets for reuse.
3. Transform & Prepare cleans and structures the data.
4. AI & Machine Learning builds and trains predictive models.
5. Visual Object displays and communicates outcomes through visualizations, dashboards, or applications.
6. Reasoning Atlas documents, contextualizes, and connects these insights within the organization's knowledge framework.

Goal: Visual Objects are the delivery layer of Reasoning Flows — transforming prepared and modeled data into interpretable, shareable insights.

Available Tools#

High Performance Computing#

Open coding environments for building custom visualizations, dashboards, REST APIs, or interactive applications.

PHP 7.4 Application

Legacy PHP environment for custom backend logic or web applications.

Best for:

• Maintaining existing PHP-based dashboards
• Simple web interfaces
• Legacy system integrations

Note: PHP 7.4 is maintained for compatibility. For new projects, use PHP 8.2.

PHP 8.2 Application

Modern PHP runtime with improved performance, security, and language features.

Best for:

• New web application development
• Custom dashboard backends
• REST API endpoints for visualization tools

Python 3.8 Advanced ML & Plotly

Python environment preconfigured with Plotly for creating interactive data visualizations.

Best for:

• Interactive charts and graphs
• Custom dashboards with Python
• Visual analytics applications
• Combining ML outputs with rich visualizations

Supported visualization libraries:

• Plotly (interactive charts)
• Matplotlib (static charts)
• Seaborn (statistical visualizations)
• Pandas plotting

Example use case:\ Build an interactive sales dashboard that allows users to filter by region, time period, and product category with drill-down capabilities.

Notebooks#

ARPIA Python Notebook

Interactive cell-based coding environment for live data analysis, visualization, and documentation. Users can write, test, and execute Python code directly in the platform while leveraging Reasoning Flows resources and datasets.

Best for:

• Exploratory data analysis with inline visualizations
• Prototyping dashboards before production deployment
• Creating visual reports with narrative documentation
• Testing and validating model outputs visually
• Sharing analysis with stakeholders

Visualization capabilities:

• Inline charts using Matplotlib, Plotly, Seaborn
• Interactive widgets for parameter exploration
• Markdown cells for documentation and context
• Export to HTML for sharing

Example workflow:

import pandas as pd
import plotly.express as px

# Load data from repository
df = repository.read('sales_gold')

# Create interactive visualization
fig = px.line(df, x='date', y='revenue', color='region',
              title='Revenue by Region Over Time')
fig.show()

Notification Engine#

AP Notification Engine

GUI-based tool for configuring and sending automated email notifications. Use this to distribute visual reports, summaries, or alerts to stakeholders.

Requirements:

• Mailgun API key (configure in platform settings)

Best for:

• Scheduled report delivery (daily, weekly, monthly)
• Alert notifications when thresholds are exceeded
• Distributing dashboard summaries to stakeholders
• Workflow completion notifications

Example use cases:

• Send a weekly sales summary email every Monday morning
• Alert the finance team when budget variance exceeds 10%
• Notify stakeholders when a new forecast is available

Prepare & Transform Tools#

AP Model Render

Executes trained models or transformations and writes structured results into tables. In the visualization context, this prepares final metrics, KPIs, or formatted datasets for display in dashboards and reports.

Best for:

• Preparing aggregated KPIs for dashboards
• Formatting model predictions for visual display
• Creating summary tables for reports
• Generating snapshot data for scheduled visualizations

Example use case:\ Run a trained sales prediction model and output a formatted table of forecasted values, ready for display in a dashboard or email report.

Web-Hook Sender#

AP Web-Hook Sender

GUI-based tool for sending webhook events or JSON payloads to external systems. Enables real-time integration between Reasoning Flows visualizations and external dashboards or applications.

Best for:

• Triggering external dashboard refreshes
• Sending data to third-party BI tools (Tableau, Power BI, etc.)
• Real-time notifications to Slack, Teams, or other platforms
• Integrating with automation tools (Zapier, Make, etc.)

Example use cases:

• Notify a Slack channel when a new report is generated
• Trigger a Power BI dataset refresh after data updates
• Send processed data to an external analytics platform

Choosing the Right Tool#

Typical Workflow Examples#

Example 1: Interactive Sales Dashboard#

sales_gold (GOLD)
    │
    ▼
Python 3.8 Advanced ML & Plotly
    • Connect to repository
    • Build Plotly dashboard
    • Deploy as web application
    │
    ▼
Interactive Dashboard (accessible via URL)

Example 2: Scheduled Email Report#

monthly_summary_gold (GOLD)
    │
    ▼
AP Model Render
    • Format KPIs and metrics
    • Generate summary table
    │
    ▼
AP Notification Engine
    • Design email template
    • Schedule weekly delivery
    │
    ▼
Stakeholder Inboxes

Example 3: Exploratory Analysis with Notebook#

raw_data (RAW) + processed_data (CLEAN)
    │
    ▼
ARPIA Python Notebook
    • Explore data patterns
    • Create visualizations
    • Document findings
    • Export to HTML
    │
    ▼
Shareable Analysis Report

Example 4: Real-Time Dashboard Integration#

predictions_optimized (OPTIMIZED)
    │
    ▼
AP Web-Hook Sender
    • Send payload to Power BI
    • Trigger dataset refresh
    │
    ▼
External BI Dashboard (updated in real-time)

Best Practices#

Choose the right tool for your audience. Use Notebooks for technical analysis and exploration. Use HPC applications for production dashboards that need to be always available.

Prepare data before visualization. Use AP Model Render or Transform & Prepare objects to format data specifically for visual display. Don't build complex data transformations inside visualization code.

Automate report delivery. Use AP Notification Engine for scheduled reports rather than manual distribution. This ensures consistency and timeliness.

Document visualizations in the Knowledge Catalog. Register important dashboards and reports as Nodes so they're discoverable and governed.

Use webhooks for real-time updates. When external dashboards need fresh data, use AP Web-Hook Sender to trigger refreshes rather than relying on polling.

Related Documentation#

• Transform & Prepare\ Learn how to structure and format data before visualization.

• AI & Machine Learning\ Explore how model outputs can be visualized and delivered through Reasoning Flows.

• Repository Tables\ Understand how to access prepared datasets for visualization.

• ARPIA Data Layer Framework\ Review which data layers are appropriate for visualization (typically GOLD and OPTIMIZED).

• Reasoning Atlas Overview\ See how visual results and KPIs are documented and connected semantically.