Our Process

ISVG’s collection methodology is straightforward. Articles are identified based on an extant list of sources, periodically reviewed and edited, and each article is read thoroughly for information pertinent to ISVG’s database. This includes information on the events, groups, individuals, and locations identified as pertinent to the incident. No information is guessed at, and all information is taken directly from the resource itself.

ISVG utilizes a human-centered, open-source entry format based on total transparency. Each source is hand-selected and entered by analysts trained in identifying information pertinent to ISVG’s mission, namely the detailed information within each event and any “linking” information contained in the article. In addition to the summary of the source that ISVG provides, the entire original source is saved, and reviewed, to prevent any information from being entered that is not backed by a publically available source.

As an open-source enterprise, ISVG utilizes only resources that are widely available to the public, and nothing that has any level of classification. In fact, the majority of ISVG’s data store comes directly from the mass media, specifically the news media available on the World Wide Web, though other sources are regularly included as well. ISVG analysts mark each source for reliability and validity, and while information may be entered from any open-source, the information identified as lower quality is discounted through the validation process until a more reliable corroborating source can be found.

In addition to the vetting process based on the analysts skills and training, ISVG has an extensive quality control system that not only assures data integrity, but prevents any data that should not be entered from remaining in the database. Specifically, every entry is reviewed by ISVG staff to make sure it meets the entry standards mentioned above.

As ISVG merely organizes existing, open-source information, it does not generate intelligence. As such, every piece of information entered into the ISVG database is reviewed both for the accuracy of the entry, as well as source appropriateness, with any non-open sourced events being immediately discarded.