USC, Annenberg School of Journalism & Communication
Annenberg Operations Analytics
Snowflake pipelines and a data-vault schema behind equipment and A/V operations.
May 2023 – May 2024
A data-vault model that tamed messy A/V and equipment operations
Equipment was going overdue and A/V and network downtime was going unmanaged because the operational data lived in disconnected systems with no shared model. I built Snowflake pipelines and a data-vault schema, chosen over a direct star schema because the sources changed shape often and data vault absorbs that without rewriting history, then automated Power BI and Streamlit reporting on top and led the weekly stakeholder reviews that turned the pain into requirements.
- Data Vault
- schema over unstable sources
- ↓ downtime
- A/V and network
- Weekly
- stakeholder reviews led
- Snowflake
- Data Vault
- Power BI
- Streamlit
- Prescriptive ML
The problem
Equipment was going overdue and A/V and network downtime was going unmanaged, because the operational data lived in disconnected systems with no shared model and no automated reporting.
What I built
- 01Built Snowflake pipelines and a data-vault schema to give the operational data a single, auditable model.
- 02Automated Power BI and Streamlit dashboards on top so reporting stopped being a manual task.
- 03Developed prescriptive ML models for staffing optimization.
- 04Streamlined staffing and CRM workflows to cut response times.
- 05Led weekly stakeholder reviews and managed server administration with a security focus.
Outcome
Reduced equipment-overdue incidents and A/V and network downtime, and cut response times through streamlined staffing and CRM workflows.