Topics, metrics, and features
System is designed to maximize precision in how evidence is captured and represented, while also ensuring that information about the same or similar things is grouped together. This is key to building and representing one system.
Words necessarily obfuscate some complexity of the system they represent. So the System Platform organizes information into three layers of precision. To make it clear what resolution you’re investigating, System color codes this as follows:
Topics make up the platform’s semantic layer. They provided shared meaning spaces to organize and explore information on System. At this highest level, System organizes metrics by the topics they measure.
Unlike metrics, topics can’t be measured in and of themselves.
Metrics on System are scientific variables used to measure topics in the world. We define metrics, simply, as anything that can be measured using a defined methodology.
Metrics like Body Mass Index (BMI) and Gross Domestic Product (GDP) help us understand our health and the economy. BMI, for example, is universally defined as the body mass divided by the square of the body height, and is expressed in units of kg/m2.
A metric may be one of several ways of measuring a topic. For example:
Depending on the goal of a study, researchers measure the topic Air Pollution in different ways. For example:
- "Toxicity of Fine Particulate Matter", when comparing levels of toxicity in various combustion sources and non-combustion sources, such as diesel engines, road dust, coal combustion, as per this study.
System also provides a way of tracking the values of metrics as a system. When metrics are connected to data with time-varying values, System displays the current value alongside the most recent previous value and the percent change between the values. When the values and recent changes are examined for systems of related metrics, you get richer context to consider the data. While depictions of changes in value can be found in many places (economic indicator reports, health records, etc.), this feature is designed to show you systems of changes in the data that are used to measure our world. Since System shows you the statistical direction of the relationships between metrics (though not necessarily the causal direction), this feature can help you predict changes in metrics before they are actually measured and consider possible actions to take to move metrics.
Features (data) are the most precise building block in System’s information architecture. Whenever possible, pieces of evidence are gathered and computed at this level to maximize accuracy and reproducibility.
System matches data, metrics, and topics as follows: A topic can be matched to any metric and a metric can be matched to any feature. This many-to-one organization allows System to aggregate statistical associations even when the underlying features and metrics are operationalized slightly differently (e.g. measured on different populations).