Observations in FRED®
What is an Observation?
In FRED®, an observation is a single data point within a time series.
Each observation consists of:
- A date
- A value
For example:
| Date | Value |
|---|---|
| 2025-01-01 | 3.4 |
| 2025-02-01 | 3.5 |
An observation represents the recorded value of an economic variable at a specific point in time.
Why Observations Matter in Economic Analysis
Observations are the atomic units of economic data.
They allow economists to:
- Measure change over time
- Compute growth rates and differences
- Detect trends, cycles, and volatility
- Perform statistical analysis
- Construct models and forecasts
All higher-level analysis is built from sequences of observations.
Components of an Observation
Each observation includes:
| Component | Description |
|---|---|
| Date | The time period associated with the value |
| Value | The recorded measurement of the series |
| Raw Value | The original value returned by the data source (e.g., “.” for missing data in FRED) |
| Derived Value | Transformed representations such as percent change or log values |
Missing Observations
FRED encodes missing data as “.”.
In RESERVE:
- “value_raw” preserves the original API response
- “value” is normalized to null
Example:
{"series_id":"CPIAUCSL","date":"2025-10-01","value":null,"value_raw":"."}Missing observations are:
- Not treated as zero
- Propagated through transformations unless explicitly handled
- Skipped in statistical computations where appropriate
This ensures transparency and avoids unintended distortion of results.
Observations vs. Series
Conceptually:
| Concept | Description |
|---|---|
| Series | A collection of observations (e.g., CPIAUCSL) |
| Observations | Individual data points within that series |
A series is composed of many observations ordered over time.
Example: Inflation Data
Consider the CPI series (CPIAUCSL).
Each monthly observation represents:
- The price level at that point in time
From these observations, economists can compute:
- Month-over-month changes
- Year-over-year inflation
- Rolling averages
- Volatility measures
Example: Volatility Analysis
To measure volatility, an analyst might:
- Compute percent changes between observations
- Apply a rolling window
- Calculate standard deviation
Each step operates directly on observations.
This highlights their role as the foundation of all transformations and analysis.
How Economists Use Observations in Practice
In workflows, observations are used to:
- Feed transformation pipelines
- Construct rolling statistics
- Align datasets across time
- Handle missing or irregular data
- Generate derived indicators
Observations are typically processed through:
- Transformations (e.g., percent change, filtering)
- Window functions (e.g., rolling averages)
- Analysis commands (e.g., summary statistics, trends)
Summary
An observation in FRED is a single time-stamped data point within a series.
It allows economists to:
- Measure economic values at specific points in time
- Build transformations and derived metrics
- Perform statistical and time-series analysis
Observations are the fundamental building blocks of all economic data analysis.