Peter Rizkallah
COLL 150 Human Development/Data Science
21 November 2021
Professor Brewer
Word Count: 1,930
Methodology Investigation
Recovering From Unexpected Bombings in Lebanon
The ammonium nitrate explosion on the port of Beirut on August 4, 2020, led to a total of 218 deaths and 7,000 injuries, accompanied by the destruction of vital health infrastructure that caused many to bleed out or treat themselves at home. According to Beirut Ammonium Nitrate Explosion: A Man-Made Disaster in Times of the COVID-19 Pandemic, “This event immediately overwhelmed the ability of Beirut hospitals, emergency medical services (EMS) agencies, first responder agencies, and other responding agencies to mount an effective response. Casualties flooded local hospitals with the less injured arriving first by private transportation or by walking to nearby hospitals and the more injured arriving later by EMS vehicles” (El Sayed, 2020). This meant that health officials in the area had no concrete plan of foreseeing this deadly event in which the current system of infrastructure was unable to support. Furthermore, many rescue teams were unable to work effectively since many of them were from outside of the country, which meant that a bulk of the recovery process was spent on informing foreign aid networks of the incident rather than having a solid and strategic plan to help people within the blast’s radius. However, Lebanese data analysts have recently implemented the use of GIS systems that account for areas in which civilians do not inhabit, such as the warehouse that stored the explosive chemicals at the port. According to Statistics Lebanon Polling and Research, “We now have listings of all dwellings, buildings, apartments, and other nonresidential uses indicated on the cluster map.” Although the use of data clusters was already part of the country’s geospatial capabilities, why were these recovery efforts not implemented, even with the use of cellular data to track affected individuals? This question could be answered with an in-depth analysis of countries that have also experienced a disaster that spread throughout their respective regions and how they were able to customize data science methods to fit their own geospatial obstacles.
Optimizing Aid by Tracking Migration with Cell Cluster Data
The Democratic Republic of the Congo is surrounded by many ancient volcanoes that erupt in an unpredictable time frame. Mount Nyiragongo is composed of volcanoes that surround the region in which people live, which could be observed to critique the current methods being used in Lebanon. Since these ancient volcanoes could not be predicted in terms of when they would erupt, scholars decided to investigate migratory patterns in an effort to effectively assign aid to the most vulnerable places, such as Goma, Bukavu, Sake, and Rutshura. Essentially, the focus of this research was aimed at discovering how to spend less time identifying which regions were impacted and more time on developing emergency responses that could be used in a sustainable way. This subscriber data was collected by Vodacom RDC and observed in their method of cluster data, taking a baseline measurement of a dataset as a percentage of subscriber mobility and comparing it to where inhabitants have relocated through a series of shaded dots on a map.
“
Figure 1: details the baseline measurement that was taken on the day of the eruption prior to the issuing of an evacuation order to the population
The method is able to forecast these migratory patterns by training the cluster data to predict where the subscribers will relocate in the event of an emergency in the future all during a 24-hour period, giving aid organizations an updated record of data to make ample time to prepare and supply materials for people who have been exposed to ash or lava.
Figure 2: details the change in subscriber movement from Goma to other routes such as the Sake and Rutshuru routes after an evacuation order was issued
On the day of the evacuation, there was a dramatic change in the shading of the Goma region, which was the most impacted by the eruption of the ancient volcanoes. The darker the shade of red was on the map, the higher the concentration of subscribers that evacuated their city. This is all relative to the subscriber’s position from the 24-hour period that started from the baseline measurement to where they relocated. Moreover, this method proved to be very reliable due to its complex adaptivity of using different baseline measurements that are based on the exact number of subscribers within every 24-hour period, meaning the measurements of mobility would be much more accurate than if they were to be held constant.
In regards to Lebanon, this could prevent people who were within the radius of an explosion from fatally bleeding out by rationing enough health kits so that national health infrastructures will not have to rely more on foreign aid. According to Eruption of Mount Nyiragongo: Estimating Population Displacement Using Mobile Operator Data by Vodacom RDC, “The Mount Nyiragongo erupted on May 22, 2021. Although the time for emergency response had passed, the estimated scale of population movements suggested that recovery from this disaster and a return to normal may require significant efforts over the coming weeks and months.” Unlike the port explosion, Congolese recovery operations were able to be guided by a concrete estimation for how long the process would take. However, disaster management in Lebanon was not reflective of this due to the emergency team’s inability to improve health infrastructures both in number and quality, which lead to the long-term hazard of ammonium nitrate becoming a toxic gas that could lead to the formation of smog and thus respiratory issues for the inhabitants of Beirut. According to A Case Study of Beirut Chemical Explosion and the Effects on Environment, “Due to no official data, we cannot estimate the exposure concentration that reaches lethal or sublethal levels in the air, soil or seawater” (Rehman, 2021). Although this explosion proved to be very devastating in terms of casualties and damage to infrastructure, it was not the first of its kind. Lebanon’s geopolitical status puts it in constant danger of being bombed, which suggests that cluster analysis should also accommodate the areas that suffer from inhaling smog and particulate matter after the initial explosion by focusing on distributing aid to communities that are within any radius of the blast. This method will help give a projection to aid officials in regards to how swiftly or long recovery efforts will be and also where the most danger is in the region.
Painting a Global Picture with Social media
In 2011, a massive tsunami struck Japan. The country’s unique archipelago island structure made many people resort to social media to describe the magnitude of destruction that was caused to their homes. The study implemented the use of network structure and community evolution on Twitter to observe the human response to the tsunami online. The data method used in this study included a topsy method that was aimed at measuring the response changes in Japanese speakers, non-native Japanese speakers, and global users. The method was used in response to the researchers’ critique on retroactive studies, which spent an excessive amount of time collecting survey and interview data much time after the disaster took place, neglecting the immediate impacts to the Japanese community. The topsy method enhances the predictive power of social media reactions by providing updated information to the public, which is needed in the state of an emergency. Its ability to group both native and global posts regarding the tsunami will prove useful in studying the allocation of aid and donations to countries that struggle with this, such as Lebanon’s disconnect with its citizens in a time of need. According to Network Structure and Community Evolution on Twitter: Human Behavior Change in Response to the 2011 Japanese Earthquake and Tsunami, “To overcome these issues, researchers have recently used more objective and timely data, generated from sensor networks such as cell phone towers, to track individual mobility and population flow for large populations in real-time, providing a unique solution for disaster response and relief management” (Lu & Brelsford, 2014).
Figure 3: depicts the data collected by the API interface using common words in each language pertaining to posts about the disaster
Each user that was identified with an “@” symbol was factored in as an individual that had the ability to mention other users on Twitter, listed as a node in the table due to the filtering of users who tweeted about the tsunami in their native language. Furthermore, this method draws connections between users and lists them as a community based on how often each user is mentioned in a post. Much like the Japanese tsunami emergency response, Lebanese health authorities were not prepared to aid the population in the aftermath of the explosion due to their focus on post-disaster as opposed to the necessary arrangements that needed to be implemented with cell cluster data, which could have been used alongside topsy methods using social media as a dataset. According to The Beirut Port Explosion: Understanding Its Impacts and How to Reduce Risks from Explosive Precursors, “The harm to the density and inherent vulnerability of the civilian population and its dependence on the web of critical interconnected services that are equally vulnerable to the damaging effects of the explosive blast” (Seddon & Shiotani, 2020). Not only could communication through the access to cellular devices have directed emergency aid officials to injured civilians, but it could have strengthened long-term support for even the aftermath of the explosion by making international donors aware of the urgent conditions sooner.
Figure 4: a visual representation of the five largest communities that formed as a result of nodes mentioning each other on Twitter, respective to each language topsy (TP-JP, TP-EN, and the Global Data Set
Comparing the Cell Cluster and Topsy Methods
To determine how these data methods should be implemented in the case of Lebanon, one must return to the central question of how geospatial capabilities can be maximized to benefit the well-being of the population. Cell cluster data is known to be used to a degree in Lebanon, but it has only just recently taken into account non-residential areas in which the explosion at the port had occurred. However, this classification placed people around these zones in danger due to their proximity to non-residential zones, causing them to be overlooked when deciding where and when to allocate aid. The modeling of the cluster method used by Vodacom RDC to monitor Mount Nyiragongo would benefit the safety of the population in Beirut by tracking the disaster over time in order to gauge a better understanding for the quantity of resources that would be needed at the time of the disaster. Since Beirut is the capital of Lebanon, it is especially important to use this method in this region because it is the epicenter of culture in the country and is heavily populated compared to other cities like Tripoli and Sidon. Furthermore, the Topsy method used to measure social media response to a disaster will supplement private organizations in Lebanon with more connections that will be able to assist in the recovery efforts of the country through donations and medical assistance, alleviating some of the pressures on health infrastructures that have already been burdened by the pandemic. Encouraging native health organizations to gain more international recognition will guide donations in the right direction and thus make entrusting these groups with the allocation of aid a lot more coherent. As evidenced by this data, these methods go hand-in-hand with each other and will domestically and internationally improve disaster management if used effectively.
Is There a Gap in These Pieces of Data Science Literature?
The development of technologies to supplement the human population in its efforts to sustain life has nonlinear potential in terms of comparing theory to its application to practical situations. In Development as Freedom, this is emphasized in the passage, “it must also take note of the empirical linkages that tie the distinct types of freedom together, strengthening their joint importance” (Sen, 1999). Although data scientists are capable of constructing simulations of events that can be empirically analyzed through the lens of experimentation, it does not negate the existence of human authority over people. If these methods are expected to be used for the greater people of Lebanon, the institutions in which they are subject must also change to accommodate them. As of now, the multiparty republic that governs Lebanon cannot reach a consensus on what should be done for the people of the country. The handling of chemicals in the port warehouse and the speed of the population’s recovery can be attributed to the way in which the government conducts itself, which has not been able to put its people first and instead has used its institutions of power for personal gains when it could be used to improve health infrastructure and methods of data collection to avoid future catastrophes.
Works Cited
El Sayed, Mazen J. “Beirut Ammonium Nitrate Explosion: A Man-Made Disaster in Times of the Covid-19 Pandemic.” Disaster Medicine and Public Health Preparedness, Cambridge University Press, 18 Nov. 2020, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7985624/. Accessed 18 Nov. 2021
Flowminder, and Vodacom RDC. “Publication: Eruption of Mount Nyiragongo: Estimating Population Displacement Using Mobile Operator Data.” Publication: Eruption of Mount Nyiragongo: Estimating Population Displacement Using Mobile Operator Data, 6 July 2021, https://www.flowminder.org/resources/publications/eruption-of-mount-nyiragongo-estimating-population-displacement-using-mobile-operator-data-call-detail-records. Accessed 18 Nov. 2021
Kumat), Sen Amartya (Amartya. Development as Freedom. Oxford University Press, 1999. Accessed 18 Nov. 2021
Lu, Xin, and Christa Brelsford. “Network Structure and Community Evolution on Twitter: Human Behavior Change in Response to the 2011 Japanese Earthquake and Tsunami.” Nature News, Nature Publishing Group, 27 Oct. 2014, https://www.nature.com/articles/srep06773. Accessed 18 Nov. 2021
Rehman, Sajid ur, et al. “Ammonium Nitrate Is a Risk for Environment: A Case Study of Beirut (Lebanon) Chemical Explosion and the Effects on Environment.” Ecotoxicology and Environmental Safety, Academic Press, 2 Jan. 2021, https://www.sciencedirect.com/science/article/pii/S0147651320316705?via%3Dihub. Accessed 18 Nov. 2021
Seddon , Bob, and Himayu Shiotani . “The Beirut Port Explosion: Understanding Its Impacts and How to Reduce Risks from Explosive Precursors.” HeinOnline, https://heinonline.org/HOL/Page?handle=hein.unl%2Fbpeui0001&id=1&collection=unl&index. Accessed 18 Nov. 2021
Statistics Lebanon, “Beyond Data: Statistics Lebanon.” Beyond Data: Statistics Lebanon Polling and Research, Stat, https://www.statisticslebanonltd.com/node/149. Accessed 18 Nov. 2021