Could CrowdOptic Be Used For Disaster Response?
Crowds—rather than sole individuals—are increasingly bearing witness to disasters large and small. Instagram users, for example, snapped 800,000 #Sandy pictures during the hurricane last year. One way to make sense of this vast volume and velocity of multimedia content—Big Data—during disasters is with PhotoSynth, as blogged here. Another perhaps more sophisticated approach would be to use CrowdOptic, which automatically zeros in on the specific location that eyewitnesses are looking at when using their smartphones to take pictures or recording videos.
“Once a crowd’s point of focus is determined, any content generated by that point of focus is automatically authenticated, and a relative significance is assigned based on CrowdOptic’s focal data attributes [...].” These include: (1) Number of Viewers; (2) Location of Focus; (3) Distance to Epicenter; (4) Cluster Timestamp, Duration; and (5) Cluster Creation, Dissipation Speed.” CrowdOptic can also be used on live streams and archival images & videos. Once a cluster is identified, the best images/videos pointing to this cluster are automatically selected.
Read full post with graphics and more links.
Data Mining Wikipedia in Real Time for Disaster Response
My colleague Fernando Diaz has continued working on an interesting Wikipedia project since he first discussed the idea with me last year. Since Wikipedia is increasingly used to crowdsource live reports on breaking news such as sudden-onset humanitarian crisis and disasters, why not mine these pages for structured information relevant to humanitarian response professionals?
In computing-speak, Sequential Update Summarization is a task that generates useful, new and timely sentence-length updates about a developing event such as a disaster. In contrast, Value Tracking tracks the value of important event-related attributes such as fatalities and financial impact. Fernando and his colleagues will be using both approaches to mine and analyze Wikipedia pages in real time. Other attributes worth tracking include injuries, number of displaced individuals, infrastructure damage and perhaps disease outbreaks. Pictures of the disaster uploaded to a given Wikipedia page may also be of interest to humanitarians, along with meta-data such as the number of edits made to a page per minute or hour and the number of unique editors.
Click on Image to Enlarge
Fernando and his colleagues have recently launched this tech challenge to apply these two advanced computing techniques to disaster response based on crowdsourced Wikipedia articles. The challenge is part of the Text Retrieval Conference (TREC), which is being held in Maryland this November. As part of this applied research and prototyping challenge, Fernando et al. plan to use the resulting summarization and value tracking from Wikipedia to verify related crisis information shared on social media. Needless to say, I’m really excited about the potential. So Fernando and I are exploring ways to ensure that the results of this challenge are appropriately transferred to the humanitarian community. Stay tuned for updates.
See also: Web App Tracks Breaking News Using Wikipedia Edits [Link]
Automatically Classifying Crowdsourced Election Reports
As part of QCRI’s Artificial Intelligence for Monitoring Elections (AIME) project, I liaised with Kaggle to work with a top notch Data Scientist to carry out a proof of concept study. As I’ve blogged in the past, crowdsourced election monitoring projects are starting to generate “Big Data” which cannot be managed or analyzed manually in real-time. Using the crowdsourced election reporting data recently collected by Uchaguzi during Kenya’s elections, we therefore set out to assess whether one could use machine learning to automatically tag user-generated reports according to topic, such as election-violence. The purpose of this post is to share the preliminary results from this innovative study, which we believe is the first of it’s kind.
Read full post with graphics.
Over 1 Million Tweets from Oklahoma Tornado Automatically Processed
My colleague Hemant Purohit at QCRI has been working with us on automatically extracting needs and offers of help posted on Twitter during disasters. When the 2-mile wide, Category 4 Tornado struck Moore, Oklahoma, he immediately began to collect relevant tweets about the Tornado’s impact and applied the algorithms he developed at QCRI to extract needs and offers of help.
Read full post
Jointly: Peer-to-Peer Disaster Recovery App
My colleague Samia Kallidis is launching a brilliant self-help app to facilitate community-based disaster recovery efforts. Samia is an MFA Candidate at the School of Visual Arts in New York. While her work on this peer-to-peer app began as part of her thesis, she has since been accepted to the NEA Studio Incubator Program to make her app a reality. NEA provides venture capital to help innovative entrepreneurs build transformational initiatives around the world. So huge congrats to Samia on this outstanding accomplishment. I was already hooked back in February when she presented her project at NYU and am even more excited now. Indeed, there are exciting synergies with the MatchApp project I’m working on with QCRI and MIT-CSAIL , which is why we’re happily exploring ways to collaborate & complement our respective initiatives.
Read full post with multiple graphics.
Humanitarianism in the Network Age: Groundbreaking Study
My colleagues at the United Nations Office for the Coordination of Humanitarian Affairs (OCHA) have just published a groundbreaking must-read study on Humanitarianism in the Network Age; an important and forward-thinking policy document on humanitarian technology and innovation. The report “imagines how a world of increasingly informed, connected and self-reliant communities will affect the delivery of humanitarian aid. Its conclusions suggest a fundamental shift in power from capital and headquarters to the people [that] aid agencies aim to assist.” The latter is an unsettling prospect for many. To be sure, Humanitarianism in the Network Age calls for “more diverse and bottom-up forms of decision-making—something that most Governments and humanitarian organizations were not designed for. Systems constructed to move information up and down hierarchies are facing a new reality where information can be generated by any-one, shared with anyone and acted by anyone.”
The purpose of this blog post (available as a PDF) is to summarize the 120-page OCHA study. In this summary, I specifically highlight the most important insights and profound implications. I also fill what I believe are some of the report’s most important gaps. I strongly recommend reading the OCHA publication in full, but if you don’t have time to leaf through the study, reading this summary will ensure that you don’t miss a beat. Unless otherwise stated, all quotes and figures below are taken directly from the OCHA report.
Read full post.
Digital Humanitarians and The Theory of Crowd Capital
An iRevolution reader very kindly pointed me to this excellent conceptual study: “The Theory of Crowd Capital”. The authors’ observations and insights resonate with me deeply given my experience in crowdsourcing digital humanitarian response. Over two years ago, I published this blog post in which I wrote that, “The value of Crisis Mapping may at times have less to do with the actual map and more with the conversations and new collaborative networks catalyzed by launching a Crisis Mapping project. Indeed, this in part explains why the Standby Volunteer Task Force (SBTF) exists in the first place.” I was not very familiar with the concept of social capital at the time, but that’s precisely what I was describing. I’ve since written extensively about the very important role that social capital plays in disaster resilience and digital humanitarian response. But I hadn’t taken the obvious next step: “Crowd Capital.”
Read full article.
Automatically Extracting Disaster-Relevant Information from Social Media
My team and I at QCRI have just had this paper (PDF) accepted at the World Wide Web (WWW 2013) conference in Rio next month. The paper relates directly to our Artificial Intelligence for Disaster Response (AIDR) project. One of our main missions at QCRI is to develop open source and freely available next generation humanitarian technologies to better manage Big (Crisis) Data. Over 20 million tweets and half-a-million Instagram pictures were posted during Hurricane Sandy, for example. In Japan, more 2,000 tweets were posted every second the day after the devastating earthquake and Tsunami struck the Eastern Coast. Recent empirical studies have shown that an important percentage of tweets posted during disaster are informative and even actionable. The challenge before is how to find those proverbial needles in the haystack and to do so in as close to real-time as possible.
So we analyzed disaster tweets posted during Hurricane Sandy (2012) and the Joplin Tornado (2011). We demonstrate that disaster-relevant information can be automatically extracted from these datasets. The results indicate that 40% to 80% of tweets that contain disaster-related information can be automatically detected. We also demonstrate that we can correctly identify the type of disaster information 80% to 90% of the time. Because these classifiers are developed using machine learning, they get more accurate with more data. This explains why we are building AIDR. Our aim is not to replace human involvement and oversight but to significantly lessen the load on humans.
Read full post.