Welcome to SnowGlobe!

SnowGlobe is a program that assists with literature searching, using a technique called snowballing. Snowballing has emerged as the gold standard for comprehensive literature reviews. The process of snowballing takes all known relevant papers and searches through their references (which papers they cite) and citations (which papers cite them). Doing this by hand is an extremely time-consuming process, and until now, there was no way to do this process in bulk. SnowGlobe has several time-saving features such as automatic duplicate removal, as well as the ability to track which titles you have already screened so you don't have to screen them multiple times.

Proceed to the Prepare Search Tab to enter the papers you'd like to search (and the papers you have already searched, if applicable).

Demos for basic searching and the running list feature are available on YouTube.

Please cite us with the following information:

McWeeny, S., Choe, J., & Norton, E. (2021). SnowGlobe: An Iterative Search Tool for Systematic Reviews and Meta-Analyses. https://doi.org/10.17605/OSF.IO/U25RN

Not Your First Search? Upload List of Previously-Screened Titles

Papers to Search



Download List of Papers to be Searched (.csv)

Search options

Set API keys

Developer tools

FAQ

Does SnowGlobe interface with Covidence/Zotero/EndNote/abstrackr etc.?

SnowGlobe does not directly interface with any reference manager, screening tool, or other organizational software for systematic reviews and meta-analyses. However, SnowGlobe does have the option to download .RIS files of your search results. SnowGlobe’s input and output formats are simple enough to easily reformat for other software.

Why Microsoft Academic (MA) and Microsoft Academic Graph (MAG)?

We chose MA and MAG for several reasons. Primarily, we chose them because they allow downloads of their entire database. This allows us to reformat, index, and extract only the essentials for quick searching. Importantly, MA and MAG also have an open-source license, which allows us to do manipulate and use the database for our own purposes. No other reference/citation tracker such as SCOPUS, Web of Science, Google Scholar, etc. have/had this capability.

Why are there slight mismatches in the search totals sometimes?

The primary way SnowGlobe searches papers is through a downloadable database that we download a few times a year. We also use an API call to MA in some of our code, and the difference between the downloaded database and the online database cause these slight mismatches.

What do you plan to do now that Microsoft Academic has announced an end date?

At the end of 2021, Microsoft Academic announced that it would be no longer available on the web. However, they would allow for one final download of their database. We will restructure our code around this, and SnowGlobe should work for the foreseeable future. However, articles written after 2021 will not be able to be searched or found using SnowGlobe. We also hope that database managers at Google Scholar, Web of Science and SCOPUS will notice the functionality of our tool and integrate it into their applications.

Is your code available?

Yes, it is available here. However, the database is password protected and most of the code depends heavily on accessing this database, so the code will not work if you launch it in R/RStudio.

SnowGlobe was created by June Choe, Jinnie Choi, and Sean McWeeny at Northwestern University. This project was made possible with the support of Elizabeth Norton and the LEARN Lab. This project was also made possible by the Open Data Commons Attribution License under which Microsoft Academic Graph falls.

If you have done snowball searching by hand, and have documentation related to how the snowball search was done, please email Sean. We are working to see how well SnowGlobe compares to hand searching.