Automatic entry and coding of social networks and dyadic peer ratings

Abstract

In small-scale communities, social-scientists can use photo-rosters to collect social network and dyadic peer-ratings data. In past work, we introduced an R package to automate photo-standardization, survey construction, and data-entry. This R package, however, lacked two key features required for fully-unsupervised data-entry. First, respondent IDs needed to be manually linked to cellphone photographs of the photo-roster before DieTryin could process the data; second, users needed to identify the locations of the photo-roster in each cellphone photograph using a point-and-click interface. To address the first shortcoming, we introduce a new Android application, DieTryinCam, which facilitates annotation of cell-phone photographs with respondent, question, and panel IDs. To address the second shortcoming, we add new functionality to the DieTryin R package, which allows for the precise location of the photograph roster to be automatically identified. Automated data entry in DieTryin now requires no user input beyond a single function call from R.