Overview

This project involved migrating ExxonMobil data from their restrictive Looker Studio environment to a new, custom-built, high-performance analytics dashboard.

The core of the transformation was the deep integration of advanced AI tools (including Augment, OpenAI Codex, and Claude Code) to build the dashboard and automate the entire deployment pipeline.

Challenge

The primary challenge was achieving absolute data parity between the existing Looker Studio dashboard and the new custom analytics. This process required the AI to not just closely model the existing data visualizations, but to exactly match the underlying analytics and metrics.

In many instances, the precise proprietary algorithm or data query powering a specific chart in the legacy system was inaccessible. This necessitated an iterative, trial-and-error approach with the AI, where we experimented with various query structures until the output perfectly replicated the established numbers and charts. This discovery process, while challenging initially, proved highly effective.

Development & Implementation

Our development strategy centered on creating a robust, AI-native deployment pipeline. This pipeline was designed to entirely automate the provisioning and deployment of the application:

  1. AI Infrastructure Creation: The system uses AI tools to define and create the necessary infrastructure on AWS servers.
  2. Infrastructure Validation: It automatically checks for the existence and health of that infrastructure.
  3. Automated Deployment: The application is then seamlessly deployed to the prepared infrastructure.

Beyond the pipeline, the new custom dashboard incorporated advanced visualizations impossible to achieve in Looker Studio, including:

  • Elaborate spider charts utilizing Likert scales.
  • Interactive charts with click-through functionality to export specific data sets.
  • The ability to double-click a visualization to instantly retrieve a list of supporting user comments for deeper context.
  • Customizable, per-chart export capabilities with user-defined columns.

Results

The successful deployment of the custom, AI-powered dashboard has generated significant time and operational savings for ExxonMobil.

Managers no longer face the manual burden of producing performance reports every week, or even multiple times a day, a necessity driven by the information limitations and slowness of the previous Looker Studio setup. Instead, clients and decision-makers can now be directed to the new custom dashboard to access up-to-date, real-time analytics with incredibly fast loading speeds.

The specialized, interactive visualizations and data context features delivered a level of data granularity and immediate actionability that fundamentally improved the client's ability to interpret and act on key metrics.

Tech Stack:

AI Tools: OpenAI Codex, Claude Code, Augment
Server Host: AWS

Features:

  • AI-Native Deployment Pipeline: Automated infrastructure creation, validation, and application deployment on AWS servers using AI tools (Augment, OpenAI Codex, Claude Code).
  • Custom Data Parity: Used AI trial-and-error to precisely match and replicate complex proprietary analytics from the legacy Looker Studio dashboards.
  • Advanced Visualizations: Development of custom charts unavailable in the legacy system, including Likert scale spider charts and enhanced interactive data displays.
  • Contextual Export Functionality: Ability to click/double-click charts to immediately export custom data columns or retrieve supporting user comments.
  • Real-Time Performance: Migration to a custom dashboard that delivers up-to-date analytics with incredibly fast loading speeds.

Achievements:

  • Significant Time Savings: Elimination of the manual, weekly (or daily) reporting burden previously required by the founders.
  • Operational Efficiency: Clients gained direct access to real-time analytics, improving decision velocity.
  • Enhanced Data Granularity: Achieved a level of data insight and immediate actionability that was impossible with the previous setup.
  • Technical Modernization: Successfully transitioned major clients off legacy tooling and onto a modern, fully automated analytics platform.