Implicit prioritization in life cycle assessment: text mining and detecting metapatterns in the literature
Life cycle assessment aims to evaluate multiple kinds of environmental impact associated with a product or process across its life cycle. Objective evaluation is a common goal, though the community recognizes that implicit valuations of diverse impacts resulting from analytical choices and choice of subject matter are present. This research evaluates whether these implicit valuations lead to detectable priority shifts in the published English language academic LCA literature over time.
A near-comprehensive investigation of the LCA literature is undertaken by applying a text mining technique known as topic modeling to over 8200 environment-related LCA journal article titles and abstracts published between 1995 and 2014.
Results and discussion
Topic modeling using MALLET software and manual validation shows that over time, the LCA literature reflects a dramatic proportional increase in attention to climate change and a corresponding decline in attention to human and ecosystem health impacts, accentuated by rapid growth of the LCA literature. This result indicates an implicit prioritization of climate over other impact categories, a field-scale trend that appears to originate mostly in the broader environmental community rather than the LCA methodological community. Reasons for proportionally increasing publication of climate-related LCA might include the relative robustness of greenhouse gas emissions as an environmental impact indicator, a correlation with funding priorities, researcher interest in supporting active policy debates, or a revealed priority on climate versus other environmental impacts in the scholarly community.
As LCA becomes more widespread, recognizing and addressing the fact that analyses are not objective becomes correspondingly more important. Given the emergence of implicit prioritizations in the LCA literature, such as the impact prioritization of climate identified here with the use of computational tools, this work recommends the development and use of techniques that make impact prioritization explicit and enable consistent analysis of result sensitivity to value judgments. Explicit prioritization can improve transparency while enabling more systematic investigation of the effects of value choices on how LCA results are used.
KeywordsEnvironmental impact Life cycle assessment Methodology Prioritization Text mining Topic model Weighting
Thanks to Mark Algee-Hewitt, Justin Grimmer, Jason Heppler, Stanford’s Environmental Assessment and Optimization group, and E-IPER students for coaching, guidance, and feedback on this project. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-114747. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.
- Blei D, Lafferty J (2006) Correlated topic models. Adv Neural Inf Proces Syst 18:147Google Scholar
- Blei DM, Ng AY, Jordan MI (2003) Latent Dirichlet allocation. J Mach Learn Res 3:993–1022Google Scholar
- Google (2011) Topic Modeling Tool. https://code.google.com/archive/p/topic-modeling-tool/
- Grubert E (2014) “Life Cycle Topic Modeling Dataset.” Stanford University. http://web.stanford.edu/~gruberte/LCA_topics/LCA_abstracts_and_titles_by_year/
- International Organization for Standardization (1996) ISO 14000: environmental management. ISO, GenevaGoogle Scholar
- International Organization for Standardization (2006) ISO 14044 International Standard. In: Environmental management–life cycle assessment–requirements and guidelines. ISO, GenevaGoogle Scholar
- Jockers M (2011) The LDA Buffet Is Now Open; Or, Latent Dirichlet Allocation for English Majors. Stanford University. https://web.stanford.edu/group/ats/cgi-bin/hivetalkin/?p=2011
- McCallum AK (2002) MALLET: A machine learning for language toolkit. http://mallet.cs.umass.edu
- Salton G (ed) (1971) The SMART retrieval system: experiments in automatic document processing. Prentice-Hall. http://jmlr.csail.mit.edu/papers/volume5/lewis04a/a11-smart-stop-list/english.stop
- Sullivan D (2001) Document warehousing and text mining: techniques for improving business operations, marketing, and sales. John Wiley & Sons, Inc., New YorkGoogle Scholar