The Accelerator has conducted research and scoping activities with scholars from around the world who study the information environment from a wide range of perspectives. Based on what we have learned from their input, we are developing tools that will help remove and reduce the most pressing barriers to high-quality and timely research.
A set of three tools – typical activity datasets, an image processing pipeline, and a screenshot research tool – are currently in development and will be launched in 2025. These tools have different utilities throughout the research process and will assist scholars in accessing, collecting, storing, and analyzing data.
Typical activity datasets will function like a weather report, for researchers to determine if particular activity they are seeing is a blip or a trend. These datasets will enable researchers to characterize what is ‘typical’ activity in various online environments to build baseline measures, and match activity with content in broadcast media.
An image processing pipeline will allow researchers to (1) perform image similarity matching, compare images within a corpora, and aggregate data at the community level, and (2) compare researcher-provided data with Accelerator-collected typical activity datasets. Currently, images are difficult for individual researchers to store and process, so this computational pipeline will convert visual media into easily usable dense embeddings and combine these embeddings with related metadata (e.g., author information).
Finally, the screenshot research tool will enable researchers to receive data directly from users who opt-in to sharing it, rather than navigating the onerous and shifting policies associated with platform-specific data access. This tool will enable a richer understanding of multimodal cross-platform content flows of screenshot-derived data from Android device users, using the most ethical and privacy-conscious methods available in the field.
We look forward to making these tools available, and developing more tools and resources that will speed and improve research on the information environment in the months and years to come.