Subside: Revenue Risk Signals

Problem

SaaS teams have usage, support, and billing data, but struggle to identify early signals of churn risk.

Solution

This project builds an end-to-end data pipeline that:

• generates synthetic SaaS data
• loads data into Google Cloud Storage (GCS)
• ingests data into BigQuery
• transforms raw data into risk signals
• visualizes results in a Streamlit dashboard

Architecture

• Data Generation: Python (subside-gen)
• Data Lake: Google Cloud Storage (GCS)
• Data Warehouse: BigQuery
• Transformations: SQL
• Visualization: Streamlit

Dashboard

The Streamlit app displays:

• Accounts by Risk Level (categorical distribution)
• Usage vs Support Activity Over Time (time series)

Dashboard Preview

How to Run

1. Generate data

subside-gen --out data_out --csv

2. Upload to GCS

gsutil cp data_out/*.csv gs://YOUR_BUCKET/

3. Load into BigQuery

bq load --autodetect --source_format=CSV ...

4. Run transformations

Execute the SQL scripts to create derived tables.

5. Run dashboard

streamlit run app/app.py

Notes

• Replace YOUR_BUCKET with your GCS bucket
• Set your GCP project ID (e.g. de2026-485318)
• Run SQL transformations in BigQuery to create derived tables