The use of social media platforms such as Twitter by affected people during crises is considered a vital source of information for crisis response. However, rapid crisis response requires real-time... Show More

The use of social media platforms such as Twitter by affected people during crises is considered a vital source of information for crisis response. However, rapid crisis response requires real-time analysis of online information. When a disaster happens, among other data processing techniques, supervised machine learning can help classify online information in real-time. However, scarcity of labeled data causes poor performance in machine training. Often labeled data from past event is available. Can past labeled data be reused to train classifiers? We study the usefulness of labeled data of past events. We observe the performance of our classifiers trained using different combinations of training sets obtained from past disasters. Moreover, we propose two approaches (target labeling and active learning) to boost classification performance of a learning scheme. We perform extensive experimentation on real crisis datasets and show the utility of past-labeled data to train machine learning classifiers to process sudden-onset crisis-related data in real-time.

In the following paper, we will present an alternate method for the detection of emotional content within social media data. Current research has presented the traditional bag-of-words method in... Show More

In the following paper, we will present an alternate method for the detection of emotional content within social media data. Current research has presented the traditional bag-of-words method in which a predefined corpus is used to measure the emotional context of each word within a message. Here we present a method in which a small subset of the data is labeled to generate a corpus which is then used to detect emotional content within the data. This research is being conducted on the dataset from hurricane Sandy in 2012. Our findings show an improvement upon the bag-of-words method. These findings would further the current research in improving the utilization of social media data within crisis response. In doing this we allow the average citizen to provide beneficial data to those in decision making roles.

Twitter is a vital source for obtaining information, especially during events such as natural disasters. Users can spread information on Twitter either by crafting new posts, which are called... Show More

Twitter is a vital source for obtaining information, especially during events such as natural disasters. Users can spread information on Twitter either by crafting new posts, which are called “tweets,” or by using the retweet mechanism to re-post previously created tweets. During natural disasters, identifying how likely a tweet is to be retweeted is crucial since it can help promote the spread of useful information in a social network such as Twitter, as well as it can help stop the spread of misinformation when corroborated with approaches that identify rumors and misinformation. In this paper, we present an analysis of retweeted tweets from two different hurricane disasters, to identify factors that affect retweetability. We then use these factors to extract features from tweets' content and user account information in order to develop models that automatically predict the retweetability of a tweet. The results of our experiments on Sandy and Patricia Hurricanes show the effectiveness of our features.

Vehicles or freight cars on fire below a bridge or inside a tunnel are exceptional events and imply difficult intervention conditions for firefighters. A buried technical network like high voltage... Show More

Vehicles or freight cars on fire below a bridge or inside a tunnel are exceptional events and imply difficult intervention conditions for firefighters. A buried technical network like high voltage electricity line, gas or steam pipeline around such a fire causes additional specifics risks. Vulnerability areas for firefighters are zones where both factors exist: a difficult incident area together with a specific risk like buried networks. They require intervention teams with specific emergency response capabilities. The paper proposes a method developed for the Paris Fire Brigade for vulnerability mapping. Results aim at improving the mobilization in allocating directly the specific responses capabilities intervention teams. Results are debated from an operational point of view. Cutting off several network lines during firefighters' interventions may strongly affect the society. In case of simultaneous incidents in vulnerable areas, firefighters could be an early warning system and inform authorities of the risk of services disruption.