riordan
commented
8 years ago I don't know why @slifty would want to talk about The World's Worst DVR, but I guess it'd be a remarkable session on creating and curating open culture for now and for future generations.
ryanpitts
[ ID ] 411ee497-d758-4b79-965b-ac2ba43546d2
[ Submitter's Name ] Daniel Schultz [ Submitter's Affiliated Organisation ] The Internet Archive [ Submitter's Twitter ] slifty
[ Space ] journalism [ Secondary Space ] digital
[ Format ] hands-on
Description
How much free air time did Donald Trump get in 2016? What was the most repeated segment of the first presidential debate? How often did Toyota air their ad for their new Prius?
This session will be about tools that allow people to ask these questions, and questions of their own.
Participants will learn to search and interact with broadcast television using the Internet Archive's corpus of fingerprinted TV programs. More specifically, folks will learn how to use open source audio fingerprinting and caption extraction tools to design and build applications that have the ability to know what exactly has been happening on TV and generate datasets based on the resulting audio and text analysis.
Agenda
This session would be somewhat (but not exclusively) code oriented -- folks would need to come with a computer and ready to play with tools. The tools (Audfprint, Duplitron 5000, and Opened Captions) will come in ready-to-use packages (npm and docker), and the audio fingerprint corpuses would be made available via USB.
There will be examples of how to use the tools, as well as a few simple projects that people can try out to get up to speed. In addition to this folks will be given time to brainstorm other ideas that come from an understanding of these tools.
Developers: will learn to use these tools and build things.
Designers: can start to design use cases enabled by these tools.
Journalists: can brainstorm story ideas and contextualized information around specific events grounded in media analysis.
Participants
Three participants: an intimate workshop where we all get set up together, discuss the possibilities offered by these technologies, and learn to use them by building out some of the examples.
Fifteen or more participants: folks will cluster into groups based on interest set (learning to use the tools / designing interfaces powered by the tools / dreaming stories answerable through the tools). If the clusters are too large, they will break down into smaller groups (5 per group).
Those who are interested in setting up the tools will be given printed instructions and thumb drives with the technologies and datasets on them. The goal of this group will be to have the participants able to write applications using the tools.
Those who are interested in designing interfaces will be given paper prototyping materials and a series of brainstorming prompts to get them thinking. The goal of this group will be to have the participants dream of applications that use the tools.
Those who are interested in the media analysis / journalism side of things will be given access to existing resources that are powered by these tools (e.g. http://politicaladarchive.org and the Internet Archive's TV Archive) and asked to explore the data and content to think of story ideas and interesting patterns.
Depending on how the session goes, at some point the groups may be mixed up so that all three groups come together and start hacking on real projects.
Outcome
Participants will walk away inspired to build new things in the real world using these open source projects. We will invite them to bring their new ideas of story ideas to newsrooms, build their own prototypes and applications out in the real world, and join the growing community of people who are discovering novel techniques around digital video.