In the past years, Thomson Reuters' calculation of the impact factor has shown several limitations. An increasing need for alternative ways of assessing the quality of scientific research is felt in the scientific community all over the world.

The Jfactor project has been set up in response to this need. A significant drive behind this project has been Thomson Reuters' denial to issue the 2012 IF of the Journal of Instrumentation (JINST – a journal owned by IOP and Sissa Medialab). The reasons behind this decision turned out to be alleged anomalies in the pattern of citations received by JINST in 2011 and stemming from proceedings published in Proc. SPIE by an author who massively cited his own work. After months of suspension, Thomson Reuters eventually agreed that JINST was entitled to receive an impact factor. This delay, however, damaged the reputation of the journal and that of its authors, not least for the lack of clarity with which the issue was handled.

The Jfactor project started at Sissa Medialab in February 2014 with data provided by INSPIRE staff, who had previously run some experimental work in 2007 – and whose comments apply to Jfactor as well. The first results were presented to the community at the beginning of April, while this web site was made available in July 2014.

For more information about Thomson Reuters IFs go to the Journal Citation Reports® (Thomson Reuters, 2014).

Technical details

Thomson Reuters impact factor and Jfactor share the same calculation method, the main difference being that Jfactor uses open data.

For example,

A = 2013 impact factor for journal X

B = citations in 2013 to articles published in journal X in 2011 and 2012

C = total number of articles published in journal X in 2011 and 2012

A = B/C

Below we describe how data are collected to compute Jfactor.

Jfactor data - collection

Publication data

We use INSPIRE's OAI-PMH interface to collect publication information. This consists of XML files that we parse and transfer to an internal DB. There are roughly 1.2 mln records.


References are collected with a heuristic process directly from the PDF files of the articles. As they are difficult to extract from the xml files, INSPIRE staff provides us with a json file as needed. This file is also parsed and the information transferred to an internal DB. There are roughly 13 mln references.

Jfactor data - sample

A table showing the latest data collected can be downloaded here.

There is a row for each year and each row contains 5 data points for each journal. Where known, the following are provided:

  1. number of publications in the previous two years from INSPIRE's data
  2. number of citations of the above in the the current year from INSPIRE's data
  3. Jfactor
  4. number of publications in the previous two years from official sources
  5. Thomson Reuters impact factor


For more information and any other requests please contact jfactor@medialab.sissa.it.

Last modified: Wed Jul 30 10:00:23 CEST 2014