What is a Decode Table?

In Decode, a Table is the fundamental building block for storing your cleaned, organized, and standardized insurance data. Think of it as a smart, structured spreadsheet specifically designed for the kind of information you work with, like premium bordereaux or claims details.

Unlike a simple spreadsheet, a Decode Table has a defined format/structure (called a schema) that you create. This structure ensures consistency across all the data you upload, making analysis much easier and more reliable.

Table Types

Decode supports two types of tables to handle different data patterns:

Time Series Tables

Time Series tables are designed for data that grows both vertically (new rows) and horizontally (new time period columns). This is ideal for data reported periodically where you track metrics over time.

• Use Cases: Premium bordereaux, claims bordereaux, policy-level financial tracking, or portfolio metrics.
• Growth Pattern: New entities (rows) can be added, and new time period columns are created automatically as data arrives. For example each row may correspond to a policy or a claim.
• Example: A "Premiums" table where each row is a policy, and each period column stores that policy's Gross Premium, Commission, and Net Premium for a given month.

Records Tables

Records tables are designed for data that grows only vertically (new rows). These are flat data tables without time-indexed columns, suitable for transaction logs or entity registries.

• Use Cases: Payment ledgers, entity registries, transaction logs, one-time data records.
• Growth Pattern: New records (rows) are added over time, but no new columns are created.
• Example: A "Payments" table where each row represents a single payment transaction with fields like Reference, Date, Amount, and Payer.

Key Characteristics of a Decode Table

• Structured Storage: Each Table corresponds to a dedicated table in the Decode database, providing robust and scalable storage.
• Defined Schema: You explicitly define the columns and their types when you create a Table. This brings order to potentially messy spreadsheet data, and help our AI powered data cleaning engine know what information to look for in your uploads.
• Flexible Growth: Time Series tables can grow both vertically and horizontally (new entities and new metric data), while Records tables grow only vertically (new entites, fixed metric data).
• Data Hub: Tables serve as the central repository from which you can inspect data directly on the Data Page or pull data into Data Functions for analysis on the Dashboard Page. Your data also allows the Decode Agent to carry out data-based tasks at your command.

Anatomy of a Table Schema

When you define a Table, you specify the following types of columns:

1. Identifier

• Purpose: This is the single most crucial column. It's the unique key that identifies each individual record (like a specific policy or claim). Decode uses this to know whether incoming spreadsheet data relates to a new record or is an update to an existing one.
• Example: Policy Number, Claim ID, Payment Reference.

2. Detail Columns

• Purpose: These columns hold information about a record that typically doesn't change when new data arrives. They describe static attributes of the entity.
• Example: For a policy, this might include Policyholder State, Class of Business, Inception Date.

3. Attribute Columns

• Purpose: These are columns you define in the Table schema but which don't necessarily exist in your uploaded spreadsheets. You assign values to these columns during the upload process using Attribute Tags and Data Profiles. They allow you to add crucial organizing context.
• Example: Wholesaler Name, Contract ID, Underwriter Initials, Region.

4. Metric Columns

• Purpose: This special structure handles the actual metrics you wish to track per row. For example you may wish to track the Recovered Salvage of a given claim per month. For Time Series tables, Decode automatically creates new columns over time as data for new periods is uploaded. Each period column stores the relevant metrics for that specific time period.
• Time Period Options: You choose the time granularity when creating a Time Series table:

  Period	Format	Example	Display
  Daily	YYYY_MM_DD	2024_01_15	Jan 15, 2024
  Monthly	YYYY_MM	2024_01	Jan 24
  Quarterly	YYYY_QN	2024_Q1	Q1 24
  Yearly	YYYY	2024	2024

• Example: If you track monthly premiums, you might have period columns like 2024_01, 2024_02, 2024_03, etc. Each of these columns would contain metrics like Gross Premium, Commission, Net Premium for that specific month.

By understanding these components, you can effectively structure your Tables to accurately reflect and organize your specific insurance data needs.

Next, let's learn how to Create a New Table.