Skip to content

Latest commit

 

History

History
102 lines (84 loc) · 4.67 KB

representing_statistics.md

File metadata and controls

102 lines (84 loc) · 4.67 KB
Raw

Representing statistics in Data Commons

Note: this document assumes familiarity with the schema.org data model.

Data Commons adopts the schema.org data model, and its schema is a superset of schema.org schema.

The most commonly used high-level schema types for representing statistical data are:

StatisticalVariable

StatisticalVariable represents any type of statistical metric that can be measured at a place and time. Some examples include: median income, median income of females, number of high school graduates, unemployment rate, prevalence of diabetes, essentially anything you might call a metric, statistic, or measure.

Consider the variable "median age of women". It is represented by a StatisticalVariable node with DCID (DC identifier) Median_Age_Person_Female in MCF format:

Node: dcid:Median_Age_Person_Female
typeOf: dcs:StatisticalVariable
populationType: schema:Person
measuredProperty: dcs:age
statType: dcs:medianValue
gender: dcs:Female

Before we get into the sematics of the above representation, as a quick primer to the MCF syntax: Node is a sentinel that defines a node in the KG, the LHS things (typeOf, measuredProperty, etc.) are DC Property entities, the RHS things (Median_Age_Person_Female, geoId/4865000, etc.) are corresponding values of those properties. Values that refer to other DC entities include namespace prefixes, such as dcid: (DC identifier), dcs: (aka DC Schema ID) and schema: (aka schema.org defined schema).

Semantically, the StatisticalVariable is modeled as a measure ("median age") on a set of things of a certain type ("people") that satisfy some set of constraints ("gender is female"). Correspondingly, StatisticalVariable types have the following properties:

  • populationType: specifies the type of thing being measured (here, Person)
  • measuredProperty: specifies the property being measured. This property is of the type defined in populationType (here, age).
  • statType (optional): specifies the type of statistic (here, medianValue). The default is measuredValue.
  • measurementQualifier (optional): additional qualifiers of the variable; e.g., "Nominal" for GDP.
  • measurementDenominator (optional): for percentages or ratios, this refers to another StatisticalVariable node.

Additionally, there can be a number of property-value (PV) pairs representing the constraints on the type identified by populationType. In this example, there is one constraint property gender (which should be of type Person) with value Female. The constraint property values are typically Enumerations (here GenderType) or Quantity nodes.

StatVarObservation

Each StatisticalVariable node is an abstract specification about which we can make observations that are grounded at a particular location and time. These is represented by the StatVarObservation type.

As an example, the statement "According to the US Census ACS 5 Year Estimates, the median age of women in San Antonio, Texas in 2014 was 34.4 years." can be represented as a StatVarObservation node in MCF format:

Node: Observation_Median_Age_Person_Female_SanAntonio_TX_2014
typeOf: dcs:StatVarObservation
variableMeasured: dcs:Median_Age_Person_Female
observationAbout: dcid:geoId/4865000
observationDate: "2014"
value: 34.4
unit: dcs:Year
measurementMethod: dcs:CensusACS5yrSurvey

The principal properties are:

  • variableMeasured: an instance of StatisticalVariable being measured
  • observationAbout: the DC entity the observation is about
  • value: numeric/string value of the observation
  • observationDate: the date of (or last day of) the observation
  • observationPeriod (optional): the length of time the observation took place
  • unit (optional): an instance of UnitOfMeasure
  • measurementMethod (optional): an Enumeration instance that describes the methodology used in the measurement (here, the ACS 5-year estimates).