Datapoints: News from the Lippincott Library of the Wharton School
R Craft: Automate your Literature Search with litsearchr
Struggling to develop a comprehensive set of terms for your literature search? Save time and effort with the R package litsearchr.
When should I use litsearchr?
This package becomes most helpful when one or more of the following describes your research:
- You have a narrow, focused topic.
- You need, potentially, many more articles.
- You’ve planned time to screen all results.
- You need a reproducible search strategy.
It means, for example, that you’re more likely to find litsearchr useful for a systematic review or meta-analysis than for a more cursory literature review. Of course, you might find other great uses for the package!
How do I set up litsearchr?
Currently, litsearcher is available from its creator’s GitHub repository. You can run this line of code in R’s console to install:
install.packages("remotes"); library(remotes); install_github("elizagrames/litsearchr", ref="main")
How can I use litsearchr in my research project?
While you can run your project in any way you’d like, litsearchr is designed to follow a roughly standard workflow:
- Get metadata (e.g.: title and abstract) from a starting set of articles relevant to your research.
- Load and prepare the article metadata.
- Extract potential keyword terms.
- Measure terms’ importance (based on co-occurrence) and set a threshold for terms to use.
- Label terms with the standardized research concepts they represent.
- Generate a Boolean search string to use in your choice of literature database(s).
For full details about litsearchr functions and techniques, see the training materials on litsearchr’s website. Penn community members also can check for live training sessions or watch a recorded workshop (R Advanced: Automate your Literature Search).
Date
October 16, 2023