Page 127 - Materials Chemistry, Second Edition

110 LIFE CYCLE ASSESSMENT HANDBOOK

etc. Therefore, it is important to seek current data. Several categories of data
are often used in inventories. Starting with the most disaggregated, these are:

• Individual process- and facility-specific: data from a particular
operation within a given facility that are not combined in any way.
• Composite: data from the same operation or activity combined
across locations.
• Aggregated: data combining more than one process operation.
• Industry-average: data derived from a representative sample of
locations and believed to statistically describe the typical opera-
tion across technologies.
• Generic: data whose representativeness may be unknown but
which are qualitatively descriptive of a process or technology.

Data can be classified by how they are created:

• Site-specific (directly measured or sampled)
• Modeled, calculated or estimated
• Non-site specific (i.e., surrogate data)
• Non-LCI data (i.e., data not originally intended for use in an LCI)
• Vendor data

Data sources are either primary or secondary:

1. Primary data come directly from the source, including:
• Interviews,
• Questionnaires or surveys,
• Bookkeeping or enterprise resource planning (ERP) system,
• Data collection tools (online, offline), and
• On-site measurements.

2. Secondary data come from reports found in:
• Databases,
• Statistics, and
• Open literature.

Unit process datasets are the basis of every LCI database and the foundation
of all LCA applications. A unit process dataset is obtained as a result of quan-
tifying inputs and outputs in relation to a quantitative reference flow from a
specific process. These inputs and outputs are generated from mathematical
relationships based on raw data that have not previously been related to the
same reference flow. An aggregated process dataset is obtained from a collec-
tion of similar unit process or other aggregated datasets. Most often, datasets
are aggregated to protect business-sensitive, competition-sensitive, or propri-
etary information, including trade secrets, patented processes, process infor-
mation used to easily derive costs, etc.