Self-serve analytics infrastructure
Built a scalable Tableau dashboard automation backend — from 10 dashboards in 2023 to 30+ in 2024, with anomaly detection and zero manual refresh.
Context
A digital marketing and base management team needed a scalable, automated reporting layer covering FIOS, FWA, and Common Customer performance. All dashboards were manual-refresh or nonexistent. Data teams were spending significant time on repetitive export and email tasks.
The problem
No daily reporting cadence. Manual refreshes. Inconsistent data access across leadership. No anomaly detection or proactive performance monitoring.
Methodology
Dashboard architecture
Established a reusable backend pattern for all dashboards:
Data model in Teradata EDW → Python/BTEQ automated refresh → Tableau extract → published to server
Designed for easy ad-hoc modification without rebuilding from scratch.
10 dashboards put into production in 2023:
Common Customer Scorecard, FIOS Scorecard, FIOS Trends, FIOS Daily Spreads, FWA Performance Book, FWA Scorecard, FWA Trends, FWA Daily Spreads, Home Performance Book, Home Churn Performance.
Automation layer (per dashboard)
- Automated data refresh via Python + BTEQ on scheduler
- Dynamic SQL query updates (date ranges, filters auto-calculated)
- Daily export pipeline:
- Cycle through filter combinations
- Export each as separate PDF
- Attach to daily email
- Auto-distribute to stakeholders
Query optimization
- Identified scripts consuming disproportionate EDW resources
- Rewrote backend logic: reduced run time, freed compute, improved refresh reliability
Documentation standard
Created step-by-step process documentation for every dashboard and automation: links, file paths, screenshots, dependencies. Enabled smooth handoff and team cross-training.
Scale-up (2024)
- Expanded Tableau server to 30+ dashboards
- Implemented anomaly detection layer for proactive performance monitoring
- Automated EMEM Executive Summary: daily consolidated performance (FIOS, FWA, Mobile) vs. targets, auto-distributed to leadership
- Spearheaded Teradata → GCP migration
Stack
Python, BTEQ, SQL / Teradata EDW, Tableau, GCP, Windows Task Scheduler.
Impact
- Zero manual refreshes across 30+ dashboards
- Daily data availability vs. prior weekly cadence
- 250+ hours saved annually (forecasting models)
- Executive leadership receiving automated daily performance summaries
- Manager: “His ability to tell a story through visualizations, automated dashboards, and insights plays a vital role in proactive churn management”