9 Understanding Social-Mediated Disaster …
165
(Smith et al. 2010) and discovered five major topics that emerged from the resolution stage of Hurricane Maria: food support, mental and physical health, fatalities,
government’s responses, and water supply. The findings will help researchers, practitioners, and decision-makers understand the public’s concerns, mitigate the public’s
uncertainty, and provide immediate disaster relief responses.
9.4 Case Study of Hurricane Maria Recovery Stage
In Fall 2017, Hurricane Maria hit multiple areas, such as Dominica, the United States
Virgin Islands, and Puerto Rico. Hurricane Maria formed on September 16, 2017,
became extratropical after September 30 and dissipated on October 2, 2017 (National
Hurricane Center 2017). By the end of August 2018, the Puerto Rico government
confirmed that estimated 2,975 people died because of Hurricane Maria (Santiago
et al. 2018). This disaster has not only caused a loss but also created uncertainty for
the public to be back to normal life.
With the purpose of understanding how the public discusses and communicates
this disaster on social media, this case study used topic modeling, latent semantic
analysis, content analysis, and word-cloud to discover the dominant topics of disaster
and risk communication surrounding Hurricane Maria during the resolution stage.
Specifically, this case study aims to answer the following question: What topics were
associated with Hurricane Maria in the resolution stages?
9.4.1 Data Collection
The tweets related to Hurricane Maria were collected with the keyword searching
function in NodeXL Professional (Smith et al. 2010), which connects to the public
free Application Programming Interface of Twitter. Specifically, by searching the
keywords “Hurricane Maria” and “HurricaneMaria,” 12,146 tweets (original tweets
and retweets) about Hurricane Maria were collected. These tweets were posted
from November 6 to November 21, 2017. Specifically, this dataset contained 7,550
vertices. This data collection strategy enables the author to cover tweets with different
hashtags and allows the case study to capture more open discussions surrounding
Hurricane Maria.
9.4.2 Data Analyses
This study utilized topic model analysis, which is a generative model providing
a probabilistic framework of text data. This case study has utilized JMP Pro13
165
(Smith et al. 2010) and discovered five major topics that emerged from the resolution stage of Hurricane Maria: food support, mental and physical health, fatalities,
government’s responses, and water supply. The findings will help researchers, practitioners, and decision-makers understand the public’s concerns, mitigate the public’s
uncertainty, and provide immediate disaster relief responses.
9.4 Case Study of Hurricane Maria Recovery Stage
In Fall 2017, Hurricane Maria hit multiple areas, such as Dominica, the United States
Virgin Islands, and Puerto Rico. Hurricane Maria formed on September 16, 2017,
became extratropical after September 30 and dissipated on October 2, 2017 (National
Hurricane Center 2017). By the end of August 2018, the Puerto Rico government
confirmed that estimated 2,975 people died because of Hurricane Maria (Santiago
et al. 2018). This disaster has not only caused a loss but also created uncertainty for
the public to be back to normal life.
With the purpose of understanding how the public discusses and communicates
this disaster on social media, this case study used topic modeling, latent semantic
analysis, content analysis, and word-cloud to discover the dominant topics of disaster
and risk communication surrounding Hurricane Maria during the resolution stage.
Specifically, this case study aims to answer the following question: What topics were
associated with Hurricane Maria in the resolution stages?
9.4.1 Data Collection
The tweets related to Hurricane Maria were collected with the keyword searching
function in NodeXL Professional (Smith et al. 2010), which connects to the public
free Application Programming Interface of Twitter. Specifically, by searching the
keywords “Hurricane Maria” and “HurricaneMaria,” 12,146 tweets (original tweets
and retweets) about Hurricane Maria were collected. These tweets were posted
from November 6 to November 21, 2017. Specifically, this dataset contained 7,550
vertices. This data collection strategy enables the author to cover tweets with different
hashtags and allows the case study to capture more open discussions surrounding
Hurricane Maria.
9.4.2 Data Analyses
This study utilized topic model analysis, which is a generative model providing
a probabilistic framework of text data. This case study has utilized JMP Pro13