Preparing Data for Integration

While Sundial can ingest any table-like format out there, some formats are better suited for data analysis compared to others. Additionally, organizations have different levels of data maturity, compliance and engineering constraints. With this in mind, Sundial recommends providing data in a few high-level formats:

1. Events - Events are granular fact data that can be ingested from your data lake or cloud warehouse. Each fact is usually identified by an event type and contains entity-level (user, transaction, order) identifiers. Examples of are app-open, clicks, views and other engagements events, subscription started and cancelleation events, etc. Each event log also contains attributes like country, platform, product-category that describe the event.
2. Backend Database Records - These are fact tables maintained by your backend applications representing orders table, transaction tables, product lists, price lists etc.
3. Pre-Aggregated Data - Sundial can also directly ingest tables that are pre-aggregated at any of the dimension levels, losing out out on entity-level data. This may be done for reducing data size, dealing with security constraints, anonimyzation of PII or to meet data-residency requirements.
4. Subscribers + Revenue Data Schema - Specifically for subscriber and revenue related information, a specific schema helps with faster integration, while maximizing analysis coverage.
5. App-Specific Data - Apps such as Stripe, RevenueCat etc may export data in very specific schemas. In such a case, Sundial can import those tables as-it-is.