Abstract
Usage telemetry can be an effective tool to assess the value and impact of an open source project, as well as enable time-strapped maintainers to prioritize support for features, versions, and/or components that are actively used. However, many open source users see incorporating usage telemetry into a project as an invasion of privacy, especially for projects that did not launch with embedded telemetry. Even with an adequate value proposition, many users will opt out, or choose not to opt in to share their data with a project, organization, or company. Tools are emerging to address this need for usage telemetry, however, many are still met with skepticism by the privacy-minded community members. It is therefore important for many projects and companies that need to show the value and impact of open source projects to have reliable proxy metrics. This proof of concept exercise seeks to form hypotheses regarding the relationship between usage as measured by embedded telemetry and as indicated by proxy metrics to evaluate whether less-invasive proxy metrics are sufficiently reliable in the detection of usage trends in open source projects.