Module 01: Core Concepts

This module introduces the key concepts you will use throughout this training. No coding is required — focus on understanding how the platform is structured and how its parts relate to each other.

The OpenDataDSL Data Model​

OpenDataDSL organises everything around master data objects. An object is a record that represents a real-world entity — a commodity, a trading location, a counterparty, or a data source. Objects carry both descriptive properties (metadata) and any number of named data items.

Object (Master Data)
├── Properties        (name, description, category, ...)
├── TimeSeries        (regular time series values)
├── SmartTimeSeries   (expression-based, calculated on demand)
├── EventTimeSeries   (built dynamically from events)
├── Curve             (a forward curve for a specific valuation date)
├── SmartCurve        (expression-based forward curve)
├── EventCurve        (forward curve built dynamically from events)
├── CurveSeries       (a time series of forward curves)
├── Matrix            (a 2D table of values)
├── Report            (a structured report linked to the object)
└── Documentation     (reference documentation attached to the object)

Services​

The platform exposes data through named services. Each service maps to a category of data and a REST endpoint. The most important services for this training are:

Active Variables​

In ODSL, you access platform services using active variable syntax:

//#region
// Fetch an object from the object service
obj = ${object:"MY.OBJECT.ID"}

// Fetch a timeseries from the data service
ts = ${data:"MY.OBJECT.ID:MY_SERIES"}

// Fetch an event list from the event service
events = ${event:"MY.OBJECT.ID:MY_EVENTS"}
//#endregion

The pattern is always ${service:"identifier"}.

Events​

An Event is a time-bounded transaction record. Unlike a timeseries observation (a single value at a point in time), an event has:

• A start time (eventstart) and optionally an end time (eventend)
• A type (defined by a schema in the type service)
• Any number of free-form properties — price, volume, currency, counterparty, etc.

Events are grouped into event lists. An event list belongs to a master data object and has a name, for example TRADES, NOMINATIONS, or OUTAGES.

The full identifier for an event list is: OBJECT_ID:EVENT_LIST_NAME

Common Uses for Events​

• Trade capture (price, volume, buyer, seller)
• Generation nominations
• Outage records
• Settlement prices with attributes
• Meter readings

TimeSeries​

A TimeSeries is a sequence of values indexed by time, with a defined calendar that controls what dates are valid. Common calendar types include:

TimeSeries are stored as properties on master data objects.

EventTimeSeries​

An EventTimeSeries is a special type of timeseries that is built dynamically from events. Instead of storing values directly, it points to an event list and a property name. When the timeseries is requested, the platform reads the events and extracts the named property as the timeseries values.

This means:

• Events are the source of truth
• The timeseries is always up to date with the latest event data
• You can filter which events contribute to the timeseries

SmartTimeSeries​

A SmartTimeSeries is an expression-based timeseries. It holds a formula rather than raw values. When requested, the formula is evaluated against one or more referenced timeseries.

This is ideal for:

• Currency conversion (multiply by an FX rate series)
• Unit conversion (multiply by a fixed factor)
• Spread calculations (series A minus series B)
• Scaled or adjusted views of data

SmartTimeSeries = BASE timeseries + expression formula + optional named inputs

Curve​

A Curve is a forward curve — a snapshot of prices or values for future delivery periods, taken on a specific valuation date (the ondate). Each entry on the curve is a Contract with a period code (e.g. M01, Q02, Y01) and a value.

Curves are stored on master data objects and versioned by ondate — each day's curve is a new version.

SmartCurve​

A SmartCurve is an expression-based forward curve. Like a SmartTimeSeries, it holds a formula evaluated against a base curve and optional named inputs rather than storing raw contract values.

Common uses:

• Spread curves (spark spread, dark spread)
• Premium or discount curves (base × factor)
• Interpolated or bootstrapped curves
• Blended curves from multiple inputs

SmartCurve = BASE curve + expression formula + optional named inputs

EventCurve​

An EventCurve is a forward curve built dynamically from events. It points to an event list and extracts one property as the contract value and another as the tenor code. When requested, the platform reads the matching events and assembles the curve on the fly.

This keeps the events as the single source of truth for curve data, mirroring how EventTimeSeries works for time series.

CurveSeries​

A CurveSeries is a timeseries with an ondate. The ondate is the date that the curve or timeseries represents — the valuation date from which all future delivery values are observed. Each value in the series represents the price of the asset at some future delivery point in time.

CurveSeries is commonly used for intraday curves with a large number of observations, such as an hourly power curve extending five years into the future. Storing this as a standard curve would be impractical due to the volume of contracts; the CurveSeries structure handles this efficiently by combining the timeseries storage model with the concept of an ondate.

Matrix​

A Matrix is a 2D table of values with labelled rows and columns. It is used for data that does not fit naturally into a timeseries or curve structure, such as:

• Volatility surfaces (strike × expiry)
• Correlation matrices
• Scenario grids

Matrices are stored on master data objects just like timeseries and curves.

Report​

A Report is a structured output — a formatted view of data — that can be attached to a master data object. Reports are generated from report configurations and can be used to publish summaries, quality snapshots, or analytical outputs linked to a specific object.

Documentation​

Documentation allows reference material to be attached directly to a master data object. This can be:

• A description of the data source or methodology
• Links to external documentation URLs
• Inline markdown content stored on the platform

This makes it easy for consumers of an object to understand what the data represents and how it is produced.

Datasets​

A Dataset is a monitoring configuration for a group of related data items. It defines:

• Which timeseries or curves are expected
• How many values are expected per delivery date
• Quality and critical check scripts to run on delivery

The platform tracks daily delivery against the dataset configuration and provides a dashboard showing completeness, quality, and any failures.

Quality Checks​

Quality checks are validation rules applied to timeseries data. They can be:

• Built-in checks — such as missing (checks for missing values), or range checks
• Script-based checks — custom ODSL functions that implement your own business rules

Quality check scripts are attached to a dataset and run automatically when new data arrives.

Summary​