We've learned that GIS is a combination of spatial and non-spatial data and we understand that non-spatial data is data about a spatial location. Within the GIS, we store non-spatial data in tables. These data tables can be viewed in ArcGIS and can actually be a few different kinds of files such as Microsoft Excel tables, Microsoft Access databases, Comma Separated Value (CSV) text files, and other kinds of text files which present as row and column based data.
Within the GIS, each spatial file has a non-spatial data table associated with it that has been given the special name of attribute table, or the data table which contains the values which could be attributed to a specific feature. Every spatial file has an attribute table, and when they are associated with a vector file, they are always visible for us to examine and perform calculations with the data. When the attribute table is associated with raster files, the attribute table may or may not be directly accessible for viewing, but understand that even if you can't open it, that doesn’t mean it doesn't exist.
Attribute tables, as we will learn in Chapter Five, contain a few mandatory and software created fields, or the columns of the table which run up and down, such as which geometry type the vector file holds and what the area of each polygon represented in the file is, but the rest of the fields are optional. These optional fields contain all the values that describe a specific feature and are stored right inside the attribute table. However, there are lots and lots of times when we need to take an external data table and join it with the vector file (as we will also learn all about in Chapter Five). These external data tables can house literally decades of information about some spatial features. If you're thinking that sounds like a lot of data, it is. It is a lot of data to store with just one spatial file, which is why we have external data tables.
3.5.2: Recognizing Data Tables in ArcCatalog
Much like vector and raster file icons, data tables have specific icons as well. Most data tables share just a few different icons while Microsoft Excel files have a special icon. ArcGIS even sees column and row presented text files (like CSVs) as table icons (other text files have a specific icon to recognize them as text files). The file icon for data tables which are not directly associated with spatial datain ArcCatalog looks much like a small data table, while Microsoft Excel files (those with a .xls or .xlsx file extension) have a similar looking table. We do not have an icon for attribute tables in ArcCatalog since we know that the advantage of looking at spatial data in the GIS is the 3 - 8 individual files are seen as one single file.
|Figure 3.14: Data Table in ArcGIS Software|
|Non-table based text files||Two example of a non-Excel data table icon||Microsoft Excel data tables|
3.5.3: Data Types in Data Tables
Nominal data (think “noun”-inal data) provides a descriptive record about data, usually the name or description of (okay, that’s an adjective but “adjective”-inal wasn’t as clever) a feature. Examples of nominal data are: Colorado, blue, pine trees. Pretty cut and dry.
Ordinal data (think “order”inal) is numeric or descriptive data which assumes an order or rank and the values are dependent on each other. While ordinal data lands along a scale or ranks values, it must also be assumed that the values do not define the scale, that is to say just because a feature has a value of 10, it is not twice a feature with a value of five. If we had an ordinal scale of soil erosion in an area where feature A. had a value of 5 and feature B. has a value of 10, we can only assume that the value of 5 falls below the value of 10, not that the scale is linear, for it may be exponential (in this case, that is not likely but the point is, ordinal values are only in order, they have no assumed scale). That is where interval/ratio come in.
Interval/ratio values are used when rank, order and absolute differences in magnitudes are being reflected. Interval/ratio values are almost always numeric, whole and positive numbers , and fall on a linear scale - even though these factors do not have to be true to be interval/ratio. Examples of ordinal data might be high, medium, and low, or soil erosion values of 1 - 10, while examples of interval/ratio data might be height, weight, depth, or temperature. As mentioned above, we cannot assume that ordinal data values have scale-dependent order or rank with any other values in the series, such as high, medium, and low; and with interval/ratio data values we can most definitely assume order, rank, and absolute differences - a man who stands 6’ tall is exactly twice as tall as a child at 3’; when it is Denver 35° on most days in January, it’s twice as warm in San Diego at 70°.