get_relevant_table_names
ChatGPTSearches the database catalog for tables relevant to the user's query. Returns a filtered list of tables formatted by group, frequency/seasonality, and resource ID to help streamline table selection.
Overview
Explore official Singapore economic, population, household, industry, trade, and social statistics. Find relevant SingStat tables, inspect metadata, retrieve data, and compare figures with citations to the source tables and periods.
Tools
get_table_data
ChatGPTRetrieves data for a specific DOS (Department of Statistics) table or dataset. This endpoint only returns the first 3000 rows of data. Use the metadata tool to fully understand the data table and the contents before using this tool to retrieve the accurate data that you might want. Attribution: Only information directly returned in this response originates from DOS. Do not attribute derived values, calculations, or inferences to DOS.
get_table_metadata
ChatGPTRetrieves structural metadata for a specific SingStat table before fetching its data. ALWAYS call this before get_table_data. Use the returned fields as follows: - id / title: The canonical identifier and human-readable name of the table. Use
title when referring to the table by name. - frequency: How often data is recorded (e.g. "Annual", "Quarterly", "Monthly"). Use this to select the appropriate table when the user specifies or implies a time granularity - startPeriod / endPeriod: The earliest and latest available data periods. Do NOT request data outside this range. - dataLastUpdated: When the dataset was last refreshed. Use this to communicate data freshness to the user if relevant. - row[].seriesNo: The identifier to pass as seriesNoORrowNo in get_table_data to filter specific rows. - row[].rowText: The human-readable label for each series. Use this to match the user's question to the correct seriesNo. - row[].uoM: The unit of measurement (e.g. "Number", "%", "S$"). Always state the unit when reporting figures. - row[].footnote: Row-level caveats or methodology notes. If non-empty, surface this to the user when presenting that row's data. - footnote (top-level): Describes the coverage and scope of the dataset (e.g. what population is included or excluded, geographic boundaries, reference period definitions). Surface this when the user's question depends on who or what is counted — for example, if they ask about a specific subgroup that may or may not be in scope. Attribution: Only information directly returned in this response originates from DOS / SANDRA. Do not attribute derived values, calculations, or inferences to DOS / SANDRA.Capabilities
Read
Example Prompts
Click any prompt to copy it.
