How does Data Buffet identify the geographic area for each time series? How do I specify the geo when picking a series? Can I retrieve multiple geos at once? How do Data Buffet geos relate to official classifications?
This content originally appeared in the April 2012 issue of Data Buffet Monthly, and has subsequently been adapted and updated.
Each time series in Data Buffet applies to a specific geographic area or location: employment in a country, retail sales in a city, shipments through a port, and so forth. But beyond that point, “where is it?” is a simple question with a complicated answer. This article reviews the high points, and links to detailed FAQ articles.
The basics
Measurement begins with individual areas (a.k.a. geographies, geos, locales, localities, spatial units, regions, territories). Commonly, a national statistical agency will establish one or more statistical geographical classifications. Usually, the areas are non-overlapping and completely cover the physical space. A complete group of similar areas constitutes a geo level (or type), and levels are stacked into a geo hierarchy. There might be alternative hierarchies for special purposes, as with Australia. A private entity can define areas for its own uses, which may not relate to the official hierarchy. Classifications are updated and new dated versions are issued. Two areas that occupy the same physical space (i.e., have the same borders) are conterminous (or coterminous).
Types of area include:
- Administrative areas found on a political map.
- Statistical areas, not necessarily consistent with the administrative areas.
- Point locations, such as air and sea ports.
- Pseudo-areas with no location, such as the “city size classes” for the U.S. CPI.
- Compounds of smaller units, such as U.S. metropolitan areas or supranational groupings.
How do we identify areas? Every area in Data Buffet has an alphanumeric geography code (a.k.a. geo code or geo). Every series is identified by a mnemonic, which consists of a concept code paired with a geo code. When possible, we model our geo codes after official ones, but with a structure that facilitates wild card expressions (see below).
- XET.DE, XET.MD, XET.PA = Multiple geo codes for one concept code: payroll employment for Delaware, Maryland, Pennsylvania
- BKP.DE, E51.DE, YPA.DE = One geo with multiple concepts: bankruptcy, payroll employment in NACIS 51 (information), personal income for Delaware
How do I see the name of an area? The name appears in Mnemonic 411 and as an optional header field in basket output.
How are areas officially identified? National and international agencies are responsible for establishing official maps, lists of areas and, in many cases, alphanumeric identifiers. For the U.S., these were formerly known as "FIPS codes" and we have borrowed the term; whenever you see the term "FIPS" or "fip" in a Data Buffet interface or output, it denotes the official identifier. You can display FIPS codes as an optional field in basket output, and search by them in Mnemonic 411, but you cannot use them as geo codes in mnemonics. Read more ...
Select and use areas
How do I pick areas? The Data Buffet catalog is a tree, and each “leaf” is a concept, which must be paired with a geo code to specify a particular series—and to do so, you pass through the Geography Wizard dialog box. You first pick a geo type, then an area within that type.
Can I retrieve multiple areas? Yes, in several different ways. With the Geo Wizard, you can pick from geo lists (see below) or multi-select several areas. When editing a basket directly, you can use wild card expressions; for example, geo codes for U.S. metropolitan divisions (2000 delineation) are uniquely identifiable as give-letter strings with first two always “DM.” (In our documentation we abbreviate this as DMaaa, meaning two fixed characters followed by three letters.) But did you want 10 MSAs, not all 300-plus? For that, use a geo list to enumerate a specific group. We provide a variety of prefab geo lists under the Tools menu, which can serve as examples to craft your own.
- One series, for a single area: FET.IUSA_MABI
- Many series, using a geo wild card: FET.IUSA_M^^^
- Many series, using a prefab geo list: FET.CBSA10_METRO_ONLY.LST
- The syntax for your custom geo list is the same: FET.MY_12_MSA.LST
Can I map data? Yes. When a concept exists over multiple geos, Data Buffet has a module for choropleth map visualization. You can pick from a wide variety of U.S. and international templates, zoom to a smaller area, scroll through time periods, and use formulas to synthesize custom data from native series.
Cast a wider net
Can I limit my search to specific areas or levels? You can create and save a filtered catalog that shrinks the catalog tree to just those concepts (and the folders that contain them) that apply to a specific geo. To check a geo level, use this trick: Pick a single representative area. Read more ...
For a given concept, what geo levels are available? There’s no general answer to this question, but we have partial solutions. You can try a filtered catalog (see above), or check the reference files for availability spreadsheets for the Global National Forecast, Global Metro Area Forecast, and U.S. ZIP code areas.
Why is coverage of this level incomplete? Sometimes a dataset simply isn't published for all possible geos within a level, and we will mention this in the "new data" article or background in Mnemonic 411. The Geo Wizard will offer only the applicable areas, wild card baskets will retrieve only matching series, and maps will show blank areas. Because scattered geo coverage complicates analysis, we produce Moody’s Analytics Estimates that fill the gaps; when browsing the catalog, the estimates will reside in a nearby folder.
- BLS CPI-U for nine MSAs: XCPIU.^^^
- Moody’s Analytics estimate for all 300-plus MSAs: XRCPIUM.IUSA_M^^^
Are there lists of geo codes and hierarchies? In the Reference Files section we have spreadsheets of global countries, regions and sub-regions, plus mappings between U.S. metro areas and their components, and of ZIP codes, counties and metro areas. Internationally, we have geo structure articles.
Do areas change? How do we respond? Geos are not static: For instance, Norway consolidates municipalities, the membership of the euro zone grows, entirely new countries are recognized by the U.N, and U.S. metro areas change yearly. Watch for Geo Change and New Geo articles in Data Buffet News. This mutability can complicate time-series analysis, but when building our subnational estimates and forecasts, we have procedures to compensate.
Do the identifiers of areas change? Of particular interest are the FIPS codes assigned to U.S. metro areas. There are situations where the composition of an area changes, but its FIPS code is retained -- however, we define a new geo code so that we can track the two compositions separately. Hence, there is not a one-to-one relationship between FIPS codes and our geo codes. Read more ...