X (formerly Twitter)

Industry
Social media
Project type
Case study
Date
March 7th, 2024

Project details

AI-powered internal analytics platform for enhanced decision-making
We partnered with Twitter to design and develop a custom, AI-enhanced internal data dashboard that empowered non-technical teams to visualize user engagement, access real-time insights, and reduce their reliance on data engineering resources.

Services

  • LangChain
  • BigQuery
  • PyTorch
  • AWS
  • Data Pipelines
  • Dashboard UI

Challenge

Twitter needed a flexible, self-service analytics tool that would allow internal teams—particularly those outside of engineering—to access usage data, explore trends, and make informed decisions without requiring SQL knowledge or analyst involvement. Their existing setup was fragmented, time-consuming, and not scalable for cross-functional insights.

Approach

We collaborated closely with product, operations, and data teams at Twitter to identify the most common reporting needs and build a centralized, easy-to-use dashboard. Our focus was on making data accessible, actionable, and visually intuitive—while integrating securely with internal infrastructure and automating routine reporting logic.

Outcome

In under 10 weeks, we built and deployed an AI-driven data pipeline and internal tooling that reduced manual processing time by 70% and accelerated feature experiment launches by 3×. The system scaled to handle millions of daily interactions without downtime.

What We Did

01/UX / UI Design of Dashboard

We designed a clean, user-friendly interface optimized for non-technical users, featuring customizable views, filterable metrics, and modular layouts tailored to each team’s needs.

02/Integration with Internal Data Systems

Connected directly to Twitter’s data pipelines and warehouse, ensuring secure, real-time data access for key user metrics, engagement trends, and performance indicators.

03/Intelligent Drill-Downs & Reporting Automation

Implemented drill-down capabilities with AI-enhanced logic that allowed users to move from high-level KPIs to detailed cohort, feature, and campaign-level insights in just a few clicks.

04/Role-Based Access & Custom User Views

Set up role-specific permissions and dashboards for different departments (e.g., Growth, Content, Ops), minimizing noise and maximizing relevance.