Be a Deciding Force in History
Help us use big science for justice:
Data science provides a deep look into factors influencing the use of force. Over the past two years, we’ve compiled data from over 8,000 newspaper, radio, and television accounts of police-protester interactions at events associated with Occupy. Because processing all of this data is a huge undertaking, we’re creating a faster, more robust method that involves machine learning and hundreds of crowd workers to extract more details. With bigger, better data on police-protester dynamics, we can run statistical analyses to show how patterns of interactions lead to violence – as well as how protests stay peaceful.
Raising $18k enables us to create open-source software to analyze big data of historic events, as well as 3 articles and 4 op-eds with lessons and insights for preventing violence.
Help us find insights for peaceful protest, such as:
- How and why police, protesters, and government officials decide to use force
- How interaction sequences escalate to violence
- How to de-escalate from violence
- When police and protesters use force strategically or as a reaction
- How the public and media can better support peaceful protest in the future
Support research that can prevent violence.
- Tell protest, police, and media groups about the Deciding Force project.
- Like our page on Facebook and follow us on Twitter .
- Back our campaign and find 3 friends to join the force!
We are Nick, Gladys, Betty, Christiana, Sofie, Jonathan, Jenny, Carly, Dan, Zach, and Fady – a team of sociologists and computer scientists at the University of California, Berkeley. Together, we believe healthy, constructive, and peaceful protest is a part of our democracy.
Tell me more!
How does my contribution support your research?
- $10k to write text classification modules for final data processing
- $2k to train crowd workers to accurately process data
- $8k for overseeing data quality control
- $8k to write 3 academic papers, 4 op-eds, and begin a book
- $2k covers perks and rewards for 'the force’ – that's you!
I’ll help you! What happens if I contribute more?
- $40k enables 5 additional academic articles and op-eds – our database has at least 20 articles worth of material in it!
- $45k supports production of algorithms from our work
- $60k helps us create a news crawler that can predict police and protester interactions based on today's news
- $100k expands our team to conduct cutting-edge news analysis like Vox.com!
Who has supported you thus far?
We are thankful for the support of the National Science Foundation (NSF) for getting us started, Berkeley's D-Lab for a wonderful environment where we develop our approach and tools, Berkeley's AMPLab for talented computer scientists generously supporting our crowd-research and machine learning work, and the Berkeley Institute for Data Science (BIDS) for Nick’s Data Science Fellowship. Now, we need your support to continue.
Tell us about your scientific independence.
We are researchers trained to use transparent processes and tools to show our work. We draw from as many sources of information as possible, and we use conservative statistical estimation strategies to let the data do the talking. Our possible bias is our working hypothesis that violence will prove to be ineffective and harmful to the legitimacy of protestors, police, and other parties most of the time. We’re reaching out because our insights are emerging and our funding is running out! Congress recently cut funding for research into politics and government . As independent researchers following high ethical standards, we are resisting sponsorship from biased parties.
What’s this open-source text analysis software all about?Big science in sociology opens a new horizon for documenting and understanding patterns of human behavior. To explain police-protester dynamics across tens of thousands of interactions and events, we’re creating advanced methods, gathering big data, and developing original software tools. We started using the time-tested method of tagging/highlighting text by categories. Then, we created our “txt_thrshr” software to engage with online workers and volunteers who help us extract, classify, and annotate information for statistical analysis. As our dataset grows, we will automate our process by training computer algorithms to extract information in the same way we do. This is called “machine learning,” and it empowers us to analyze human behaviours in their intricate contexts and weaving sequences. This research ultimately can produce findings and information enabling better decisions by protester, police, the public and the media. To learn more, nerd-out on this paper – “Researchers to Crowds to Algorithms: Building Large, Complex, and Transparent Databases in the Age of Data Science ” – and check out www.textthresher.org .