The blog of the John S. and James L. Knight Foundation
Liz Eddy is director of communications and events and Bob Filbin is chief data scientist at Crisis Text Line, a text messaging support service for people in crisis originally developed with Knight Foundation support. Knight is awarding Crisis Text Line, a previous winner of the Knight News Challenge, an additional $1.4 million in support as part of a $7 million funding round.
Imagine you are a principal at a school: Do you know what is the worst time of day for bullying? Do you know what is the worst day of the week for eating disorders?
Crisis Text Line has this data. And more.
Think about the massive juicy data set Crisis Text Line has collected. More than 9 million messages have been exchanged on this free 24/7 text line for people in crisis in just over two years. The volume, velocity and variety of these messages make this data set a very interesting corpus. Better yet? It’s real time.
The National Institutes of Health and Centers for Disease Control and Prevention might have more data, but it takes years to collect, store and analyze. The Crisis Text Line data set is auto-tagged in real time and made available within seconds. So, that principal is looking at trends right now, with current students, teachers, vendors, social media, etc. Finally, school data can match the pace of school challenges.
Same for police departments. Or, policymakers. Or, academics and researchers.
Simply, Crisis Text Line is committed to leveraging its data to inform people and communities. For too long, this information has been slow, expensive and hard to consume. We believe a strong democracy depends on open data.
You can see our v1 sharing at CrisisTrends.org right now. This work was funded by Knight Foundation. And, it’s just the beginning. In 2016, we plan to launch v2 of CrisisTrends.org and open up two other levels of data.
Currently, our Level 1 data can be found on Crisis Trends, which provides aggregate, real-time crisis data to the public. You can find data on the top 20 issues experienced by people who contact Crisis Text Line and segment it by hour of day, day of week, month over month, or by state. We also show frequency word clouds of the top 50 words used by texters for each issue. For example, for family issues, “mom” is mentioned much more frequently than “dad.”
Since the launch of Crisis Trends, we have had many requests from students, journalists, policymakers and beyond who are interested in seeing additional variables. As our data corpus grows in 2016, visitors to the website will be able to ask more specific questions. For example, queryable regions will shrink from state to area code to ZIP code, and data on individual days, such as Christmas, will become available. You will also be able to see the top words associated with specific issues; for example, what words do people being bullied use to describe their experiences and their aggressors? These specific types of insights are what many journalists and researchers seek, and they will be provided to communities and policymakers with more context to take action.
By the end of 2016, we expect Crisis Text Line to process 100,000 messages per day and to have a corpus of 40 million messages. This volume will allow Crisis Text Line to continue to be one of the largest public health data sets in the world. And, the only real-time data set on crisis of this size in the United States. Imagine the possibilities.