Tag Archives: Situational

Did Terrorists Use Twitter to Increase Situational Awareness?

Those who are still skeptical about the value of Twitter for real-time situational awareness during a crisis ought to ask why terrorists likely think otherwise. In 2008, terrorists carried out multiple attacks on Mumbai in what many refer to as the worst terrorist incident in Indian history. This study, summarized below, explains how the terrorists in question could have used social media for coor-dination and decision-making purposes.

The study argues that “the situational information which was broadcast through live media and Twitter contributed to the terrorists’ decision making process and, as a result, it enhanced the effectiveness of hand-held weapons to accomplish their terrorist goal.” To be sure, the “sharing of real time situational information on the move can enable the ‘sophisticated usage of the most primitive weapons.'” In sum, “unregulated real time Twitter postings can contribute to increase the level of situation awareness for terrorist groups to make their attack decision.”

According to the study, “an analysis of satellite phone conversations between terrorist commandos in Mumbai and remote handlers in Pakistan shows that the remote handlers in Pakistan were monitoring the situation in Mumbai through live media, and delivered specific and situational attack commands through satellite phones to field terrorists in Mumbai.” These conversations provide “evidence that the Mumbai terrorist groups understood the value of up-to-date situation information during the terrorist operation. [...] They under-stood that the loss of information superiority can compromise their operational goal.”

Handler: See, the media is saying that you guys are now in room no. 360 or 361. How did they come to know the room you guys are in?…Is there a camera installed there? Switch off all the lights…If you spot a camera, fire on it…see, they should not know at any cost how many of you are in the hotel, what condition you are in, where you are, things like that… these will compromise your security and also our operation […]

Terrorist: I don’t know how it happened…I can’t see a camera anywhere.

A subsequent phone conversation reveals that “the terrorists group used the web search engine to increase their decision making quality by employing the search engine as a complement to live TV which does not provide detailed information of specific hostages. For instance, to make a decision if they need to kill a hostage who was residing in the Taj hotel, a field attacker reported the identity of a hostage to the remote controller, and a remote controller used a search engine to obtain the detailed information about him.”

Terrorist: He is saying his full name is K.R.Ramamoorthy.

Handler: K.R. Ramamoorthy. Who is he? … A designer … A professor … Yes, yes, I got it …[The caller was doing an internet search on the name, and a results showed up a picture of Ramamoorthy] … Okay, is he wearing glasses? [The caller wanted to match the image on his computer with the man before the terrorists.]

Terrorist: He is not wearing glasses. Hey, … where are your glasses?

Handler: … Is he bald from the front?

Terrorist: Yes, he is bald from the front …

The terrorist group had three specific political agendas: “(1) an anti-India agenda, (2) an anti-Israel and anti-Jewish agenda, and (3) an anti-US and anti-Nato agenda.” A content analysis of 900+ tweets posted during the attacks reveal whether said tweets may have provided situational awareness information in support of these three political goals. The results: 18% of tweets contained “situa-tional information which can be helpful for Mumbai terrorist groups to make an operational decision of achieving their Anti-India political agenda. Also, 11.34% and 4.6% of posts contained operationally sensitive information which may help terrorist groups to make an operational decision of achieving their political goals of Anti-Israel/Anti-Jewish and Anti-US/Anti-Nato respectively.”

In addition, the content analysis found that “Twitter site played a significant role in relaying situational information to the mainstream media, which was monitored by Mumbai terrorists. Therefore, we conclude that the Mumbai Twitter page in-directly contributed to enhancing the situational awareness level of Mumbai terrorists, although we cannot exclude the possibility of its direct contribution as well.”

In conclusion, the study stresses the importance analyzing a terrorist group’s political goals in order to develop an appropriate information control strategy. “Because terrorists’ political goals function as interpretative filters to process situational information, understanding of adversaries’ political goals may reduce costs for security operation teams to monitor and decide which tweets need to be controlled.”

bio

See also: Analyzing Tweets Posted During Mumbai Terrorist Attacks [Link]

Debating the Value of Tweets For Disaster Response (Intelligently)

With every new tweeted disaster comes the same old question: what is the added value of tweets for disaster response? Only a handful of data-driven studies actually bother to move the debate beyond anecdotes. It is thus high time that a meta-level empirical analysis of the existing evidence be conducted. Only then can we move towards a less superficial debate on the use of social media for disaster response and emergency management.

In her doctoral research Dr. Sarah Vieweg found that between 8% and 24% of disaster tweets she studied “contain information that provides tactical, action-able information that can aid people in making decisions, advise others on how to obtain specific information from various sources, or offer immediate post-impact help to those affected by the mass emergency.” Two of the disaster datasets that Vieweg analyzed were the Red River Floods of 2009 and 2010. The tweets from the 2010 disaster resulted in a small increase of actionable tweets (from ~8% to ~9%). Perhaps Twitter users are becoming more adept at using Twitter during crises? The lowest number of actionable tweets came from the Red River Floods of 2009, whereas the highest came from the Haiti Earthquake of 2010. Again, there is variation—this time over space.

In this separate study, over 64,000 tweets generated during Thailand’s major floods in 2011 were analyzed. The results indicate that about 39% of these tweets belonged to the “Situational Awareness and Alerts” category. “Twitter messages in this category include up-to-date situational and location-based information related to the flood such as water levels, traffic conditions and road conditions in certain areas. In addition, emergency warnings from authorities advising citizens to evacuate areas, seek shelter or take other protective measures are also included.” About 8% of all tweets (over 5,000 unique tweets) were “Requests for Assistance,” while 5% were “Requests for Information Categories.”

TwitterDistribution

In this more recent study, researchers mapped flood-related tweets and found a close match between that resulting map and the official government flood map. In the map below, tweets were normalized, such that values greater than one mean more tweets than would be expected in normal Twitter traffic. “Unlike many maps of online phenomena, careful analysis and mapping of Twitter data does NOT simply mirror population densities. Instead con-centration of twitter activity (in this case tweets containing the keyword flood) seem to closely reflect the actual locations of floods and flood alerts even when we simply look at the total counts.” This also implies that a relatively high number of flood-related tweets must have contained accurate information.

TwitterMapUKfloods

Shifting from floods to fires, this earlier research analyzed some 1,700 tweets generated during Australia’s worst bushfire in history. About 65% of the tweets had “factual details,” i.e., “more than three of every five tweets had useful infor-mation.” In addition, “Almost 22% of the tweets had geographical data thus identifying location of the incident which is critical in crisis reporting.” Around 7% of the tweets were seeking information, help or answers. Finally, close to 5% (about 80 tweets) were considered “directly actionable.”

Preliminary findings from applied research that I am carrying out with my Crisis Computing team at QCRI also reveal variation in value. In one disaster dataset we studied, up to 56% of the tweets were found to be informative. But in two other datasets, we found the number of informative tweets to be very low. Meanwhile, a recent Pew Research study found that 34% of tweets during Hurricane Sandy “involved news organizations providing content, government sources offering information, people sharing their own eyewitness accounts and still more passing along information posted by others.” In addition, “fully 25% [of tweets] involved people sharing photos and videos,” thus indicating “the degree to which visuals have become a more common element of this realm.”

Finally, this recent study analyzed over 35 million tweets posted by ~8 million users based on current trending topics. From this data, the authors identified 14 major events reflected in the tweets. These included the UK riots, Libya crisis, Virginia earthquake and Hurricane Irene, for example. The authors found that “on average, 30% of the content about an event, provides situational awareness information about the event, while 14% was spam.”

So what can we conclude from these few studies? Simply that the value of tweets for disaster response can vary considerably over time and space. The debate should thus not center around whether tweets yield added value for disaster response but rather what drives this variation in value. Identifying these drivers may enable those with influence to incentivize high-value tweets.

This interesting study, “Do All Birds Tweet the Same? Characterizing Twitter Around the World,” reveals some very interesting drivers. The social network analysis (SNA) of some 5 million users and 5 billion tweets across 10 countries reveals that “users in the US give Twitter a more informative purpose, which is reflected in more globalized communities, which are more hierarchical.” The study is available here (PDF). This American penchant for posting “informative” tweets is obviously not universal. To this end, studying network typologies on Twitter may yield further insights on how certain networks can be induced—at a structural level—to post more informative tweets following major disasters.

Twitter Pablo Gov

Regardless of network typology, however, policy still has an important role to play in incentivizing high-value tweets. To be sure, if demand for such tweets is not encouraged, why would supply follow? Take the forward-thinking approach by the Government of the Philippines, for example. The government actively en-couraged users to use specific hashtags for disaster tweets days before Typhoon Pablo made landfall. To make this kind of disaster reporting via twitter more actionable, the Government could also encourage the posting of pictures and the use of a structured reporting syntax—perhaps a simplified version of the Tweak the Tweet approach. Doing so would not only provide the government with greater situational awareness, it would also facilitate self-organized disaster response initiatives.

In closing, perhaps we ought to keep in mind that even if only, say, 0.001% of the 20 million+ tweets generated during the first five days of Hurricane Sandy were actionable and only half of these were accurate, this would still mean over a thousand informative, real-time tweets, or about 15,000 words, or 25 pages of single-space, relevant, actionable and timely disaster information.

PS. While the credibility and veracity of tweets is an important and related topic of conversation, I have already written at length about this.

Situational Awareness in Mass Emergency: Behavioral & Linguistic Analysis of Disaster Tweets

Sarah Vieweg‘s doctoral dissertation from the University of Colorado is a must-read for anyone interested in the use of twitter during crises. I read the entire 300-page study because it provides important insights on how automated natural language processing (NLP) can be applied to the Twittersphere to provide situational awareness following a sudden-onset emergency. Big thanks to Sarah for sharing her dissertation with QCRI. I include some excerpts below to highlight the most important findings from her excellent research.

Introduction

“In their research on human behavior in disaster, Fritz and Marks (1954) state: ‘[T]he immediate problem in a disaster situation is neither un-controlled behavior nor intense emotional reaction, but deficiencies of coordination and organization, complicated by people acting upon individual…definitions of the situation.'”

“Fritz and Marks’ assertion that people define disasters individually, which can lead to problematic outcomes, speaks to the need for common situational awareness among affected populations. Complete information is not attained during mass emergency, else it would not be a mass emergency. However, the more information people have and the better their situational awareness, and the better equipped they are to make tactical, strategic decisions.”

“[D]uring crises, people seek information from multiple sources in an attempt to make locally optimal decisions within given time constraints. The first objective, then, is to identify what tweets that contribute to situational awareness ‘look like’—i.e. what specific information do they contain? This leads to the next objective, which is to identify how information is communicated at a linguistic level. This process provides the foundation for tools that can automatically extract pertinent, valuable information—training machines to correctly ‘understand’ human language involves the identification of the words people use to communicate via Twitter when faced with a disaster situation.”

Research Design & Results

Just how much situational awareness can be extracted from twitter during a crisis? What constitutes situational awareness in the first place vis-a-vis emergency response? And can the answer to these questions yield a dedicated ontology that can be fed into automated natural language processing platforms to generate real-time, shared awareness? To answer these questions, Sarah analyzed four emergency events: Oklahoma Fires (2009), Red River Floods (2009 & 2010) and the Haiti Earthquake (2010).

She collected tweets generated during each of these emergencies and developed a three-step qualitative coding process to analyze what kinds of information on Twitter contribute to situational awareness during a major emergency. As a first step, each tweet was categorized as either:

O: Off-topic
“Tweets do not contain any information that mentions or relates to the emergency event.”

R: On-topic and Relevant to Situational Awareness
“Tweets contain information that provides tactical, actionable information that can aid people in making decisions, advise others on how to obtain specific information from various sources, or offer immediate post- impact help to those affected by the mass emergency.”

N: On-topic and Not Relevant to Situational Awareness
“Tweets are on-topic because they mention the emergency by including offers of prayer and support in relation to the emergency, solicitations for donations to charities, or casual reference to the emergency event. But these tweets do not meet the above criteria for situational relevance.”

The O, R, and N coding of the crisis datasets resulted in the following statistics for each of the four datasets:

For the second coding step, on-topic relevant tweets were annotated with more specific information based on the following coding rule:

S: Social Environment
“These tweets include information about how people and/or animals are affected by a hazard, questions asked in relation to the hazard, responses to the hazard and actions to take that directly relate to the hazard and the emergency situation it causes. These tweets all include description of a human element in that they explain or display human behavior.”

B: Built Environment
“Tweets that include information about the effect of the hazard on the built environment, including updates on the state of infrastructure, such as road closures or bridge outages, damage to property, lack of damage to property and the overall state or condition of structures.”

P: Physical Environment
“Tweets that contain specific information about the hazard including particular locations of the hazard agent or where the hazard agent is expected or predicted to travel or predicted states of the hazard agent going forward, notes about past hazards that compare to the current hazard, and how weather may affect hazard conditions. These tweets additionally include information about the type of hazard in general [...]. This category also subsumes any general information about the area under threat or in the midst of an emergency [...].”

The result of this coding for Haiti is depicted in the figures below.

According to the results, the social environment (‘S’) category is most common in each of the datasets. “Disasters are social events; in each disaster studied in this dissertation, the disaster occurred because a natural hazard impacted a large number of people.”

For the third coding step, Sarah created a comprehensive list of several dozen  “Information Types” for each “Environment” using inductive, data-driven analysis of twitter communications, which she combined with findings from the disaster literature and official government procedures for disaster response. In total, Sarah identified 32 specific types of information that contribute to situational awareness. The table below compares the Twitter Information Types for all three environments as related to government procedures, for example.

“Based on the discourse analysis of Twitter communications broadcast during four mass emergency events,” Sarah identified 32 specific types of information that “contribute to situational awareness. Subsequent analysis of the sociology of disaster literature, government documents and additional research on the use of Twitter in mass emergency uncovered three additional types of information.”

In sum, “[t]he comparison of the information types [she] uncovered in [her] analysis of Twitter communications to sociological research on disaster situations, and to governmental procedures, serves as a way to gauge the validity of [her] ground-up, inductive analysis.” Indeed, this enabled Sarah to identify areas of overlap as well as gaps that needed to be filled. The final Information Type framework is listed below:

And here are the results of this coding framework when applied to the Haiti data:

“Across all four datasets, the top three types of information Twitter users communicated comprise between 36.7-52.8% of the entire dataset. This is an indication that though Twitter users communicate about a variety of informa-tion, a large portion of their attention is focused on only a few types of in-formation, which differ across each emergency event. The maximum number of information types communicated during an event is twenty-nine, which was during the Haiti earthquake.”

Natural Language Processing & Findings

The coding described above was all done manually by Sarah and research colleagues. But could the ontology she has developed (Information Types) be used to automatically identify tweets that are both on-topic and relevant for situational awareness? To find out, she carried out a study using VerbNet.

“The goal of identifying verbs used in tweets that convey information relevant to situational awareness is to provide a resource that demonstrates which VerbNet classes indicate information relevant to situational awareness. The VerbNet class information can serve as a linguistic feature that provides a classifier with information to identify tweets that contain situational awareness information. VerbNet classes are useful because the classes provide a list of verbs that may not be present in any of the Twitter data I examined, but which may be used to describe similar information in unseen data. In other words, if a particular VerbNet class is relevant to situational awareness, and a classifier identifies a verb in that class that is used in a previously unseen tweet, then that tweet is more likely to be identified as containing situational awareness information.”

Sarah identified 195 verbs that mapped to her Information Types described earlier. The results of using this verb-based ontology are mixed, however. “A majority of tweets do not contain one of the verbs in the identified VerbNet classes, which indicates that additional features are necessary to classify tweets according to the social, built or physical environment.”

However, when applying the 195 verbs to identify on-topic tweets relevant to situational awareness to previously unused Haiti data, Sarah found that using her customized VerbNet ontology resulted in finding 9% more tweets than when using her “Information Types” ontology. In sum, the results show that “using VerbNet classes as a feature is encouraging, but other features are needed to identify tweets that contain situational awareness information, as not all tweets that contain situational awareness information use one of the verb members in the […] identified VerbNet classes. In addition, more research in this area will involve using the semantic and syntactic information contained in each VerbNet class to identify event participants, which can lead to more fine-grained categorization of tweets.”

Conclusion

“Many tweets that communicate situational awareness information do not contain one of the verbs in the identified VerbNet classes, [but] the information provided with named entities and semantic roles can serve as features that classifiers can use to identify situational awareness information in the absence of such a verb. In addition, for tweets correctly identified as containing information relevant to situational awareness, named entities and semantic roles can provide classifiers with additional information to classify these tweets into the social, built and physical environment categories, and into specific information type categories.”

“Finding the best approach toward the automatic identification of situational awareness information communicated in tweets is a task that will involve further training and testing of classifiers.”