Research Paper
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 1/9
Title:
Database:
Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Humanitarian Operations. By: Swaminathan, Jayashankar M., Production & Operations Management, 10591478, Sep2018, Vol. 27, Issue 9
Business Source Premier
Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Humanitarian Operations
There has been a significant increase in the scale and scope of humanitarian efforts over the last decade. Humanitarian operations need to be—rapid, impactful, sustained, and efficient (RISE). Big data offers many opportunities to enable RISE humanitarian operations. In this study, we introduce the role of big data in humanitarian settings and discuss data streams which could be utilized to develop descriptive, prescriptive, and predictive models to significantly impact the lives of people in need.
big data; humanitarian operations; analytics
Introduction Humanitarian efforts are increasing on a daily basis both in terms of scale and scope. This past year has been terrible in terms of devastations and losses during hurricanes and earthquake in North America. Hurricanes Harvey and Irma are expected to lead to losses of more than $150 billion US dollars due to damages and lost productivity (Dillow [ 8] ). In addition, more than 200 lives have been lost and millions of people have suffered from power outages and shortage of basic necessities for an extended period of time in the United States and the Caribbean. In the same year, a 7.1 earthquake rattled Mexico City killing more than 150 people and leaving thousands struggling to get their lives back to normalcy (Buchanan et al. [ 2] ). Based on the Intergovernmental Panel on Climate Change, NASA predicts that global warming could possibly lead to increase in natural calamities such as drought, intensity of storms, hurricanes, monsoons, and mid‐latitude storms in the upcoming years. Simultaneously, the geo‐political, social, and economic tensions have increased the need for humanitarian operations globally; such impacts have been experienced due to the crisis in Middle East, refugees in Europe, the systemic needs related to drought, hunger, disease, and poverty in the developing world, and the increased frequency of random acts of terrorism. According to the Global Humanitarian Assistance Report, 164.2 million people across 46 countries needed some form of humanitarian assistance in 2016 and 65.6 million people were displaced from their homes, the highest number witnessed thus far. At the same time, the international humanitarian aid increased to all time high of $27.3 billion US dollars from $16.1 billion US dollars in 2012. Despite that increase, common belief is that
Listen American Accent
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 2/9
funding is not sufficient to meet the growing humanitarian needs. Therefore, humanitarian organizations will continue to operate under capacity constraints and will need to innovate their operations to make them more efficient and responsive.
There are many areas in which humanitarian operations can improve. Humanitarian operations are often blamed for being slow or unresponsive. For example, the most recent relief efforts for Puerto Rico have been criticized for slow response. These organizations also face challenges in being able to sustain a policy or best practice for an extended period of time because of constant turnover in personnel. They are often blamed for being inefficient in how they utilize resources (Vanrooyen [ 29] ). Some of the reasons that contribute to their inefficiency include operating environment such as infrastructure deficiencies in the last mile, socio‐political tensions, uncertainty in funding, randomness of events and presence of multiple agencies and stake holders. However, it is critical that humanitarian operations show high level of performance so that every dollar that is routed in these activities is utilized to have the maximum impact on the people in need. Twenty‐one donor governments and 16 agencies have pledged at the World Humanitarian Summit in 2016 to find at least one billion USD in savings by working more efficiently over the next 5 years (Rowling [ 24] ).
We believe the best performing humanitarian operations need to have the following characteristics—they need to be Rapid, they have to be Impactful in terms of saving human lives, should be effective in terms of providing Sustained benefits and they should be highly Efficient. We coin RISE as an acronym that succinctly describes the characteristics of successful humanitarian operations and it stands for Rapid, Impactful, Sustained, and Efficient.
One of the major opportunities for improving humanitarian operations lies in how data and information are leveraged to develop above competencies. Traditionally, humanitarian operations have suffered from lack of consistent data and information (Starr and Van Wassennhove [ 26] ). In these settings, information comes from a diverse set of stakeholders and a common information technology is not readily deployable in remote parts of the world. However, the Big Data wave that is sweeping through all business environments is starting to have an impact in humanitarian operations as well. For example, after the 2010 Haiti Earthquake, population displacement was studied for a period of 341 days using data from mobile phone and SIM card tracking using FlowMinder. The data analysis allowed researchers to predict refugee locations 3 months out with 85% accuracy. This analysis facilitated the identification of cholera outbreak areas (Lu et al. [ 18] ). Similarly, during the Typhoon Pablo in 2012, the first official crisis map was created using social media data that gave situation reports on housing, infrastructure, crop damage, and population displacement using metadata from Twitter. The map became influential in guiding both UN and Philippines government agencies (Meier [ 21] ).
Big Data is defined as large volume of structured and unstructured data. The three V's of Big Data are Volume, Variety, and Velocity (McCafee and Brynjolfsson [ 19] ). Big Data Analytics examines large amounts of data to uncover hidden patterns and correlations which can then be utilized to develop intelligence around the operating environment to make better decisions. Our goal in this article is to lay out a framework and present examples around how Big Data Analytics could enable RISE humanitarian operations.
Humanitarian Operations—Planning and Execution
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 3/9
Planning and Execution are critical aspects of humanitarian operations that deal with emergencies (like hurricanes) and systemic needs (hunger). All humanitarian operations have activities during preparedness phase (before) and disaster phase (during). Emergencies also need additional focus on the recovery phase (after). Planning and execution decisions revolve around Where, When, How, and What. We will take the UNICEF RUTF supply chain for the Horn of Africa (Kenya, Ethiopia, and Somalia) as an example. RUTF (ready to use therapeutic food) also called Plumpy’ Nut is a packaged protein supplement that can be given to malnourished children under the age of 5 years. The supplement was found to be very effective; therefore, the demand for RUTF skyrocketed, and UNICEF supply chain became over stretched (Swaminathan [ 27] ). UNICEF supply chain showed many inefficiencies due to long lead times, high transportation costs, product shortages, funding uncertainties, severe production capacity constraints, and government regulations (So and Swaminathan [ 25] ). Our analysis using forecasted demand data from the region found that it was important to determine where inventory should be prepositioned (in Kenya or in Dubai). The decision greatly influenced the speed and efficiency of distribution of RUTF. The amount of prepositioned inventory also needed to be appropriately computed and operationalized (Swaminathan et al. [ 28] ). Given that the amount of funding and timing showed a lot of uncertainty, when funding was obtained, and how inventory was procured and allocated, dramatically influenced the overall performance (Natarajan and Swaminathan [ 22] , [ 23] ). Finally, understanding the major roadblocks to execution and addressing those for a sustained solution had a great impact on the overall performance. In the UNICEF example, solving the production bottleneck in France was critical. UNICEF was able to successfully diversify its global supply base and bring in more local suppliers into the network. Along with the other changes that were incorporated, UNICEF RUTF supply chain came closer to being a RISE humanitarian operations and estimated that an additional one million malnourished children were fed RUTF over the next 5 years (Komrska et al. [ 15] ). There are a number of other studies that have developed robust optimization models and analyzed humanitarian settings along many dimensions. While not an exhaustive list, these areas include humanitarian transportation planning (Gralla et al. [ 13] ), vehicle procurement and allocation (Eftekar et al. [ 9] ), equity and fairness in delivery (McCoy and Lee [ 20] ), funding processes and stock‐outs (Gallien et al. [ 12] ), post‐disaster debris operation (Lorca et al. [ 17] ), capacity planning (Deo et al. [ 6] ), efficiency drivers in global health (Berenguer et al. [ 1] ), and decentralized decision‐making (Deo and Sohoni [ 5] ). In a humanitarian setting, the following types of questions need to be answered.
Where
a.Where is the affected population? Where did it originate? Where is it moving to?
b.Where is supply going to be stored? Where is the supply coming from? Where will the distribution points be located?
c.Where is the location of source of disruption (e.g., hurricane)? Where is it coming from? Where is moving to?
d.Where are the debris concentrated after the event?
When
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 4/9
a.When is the landfall or damage likely to occur?
b.When is the right time to alert the affected population to minimize damages as well as unwanted stress?
c.When should delivery vehicles be dispatched to the affected area?
d.When should supply be reordered to avoid stock‐outs or long delays?
e.When should debris collection start?
How
a.How should critical resources be allocated to the affected population?
b.How much of the resources should be prepositioned?
c.How many suppliers or providers should be in the network?
d.How to transport much needed suppliers and personnel in the affected areas?
e.How should the affected population be routed?
What
a.What types of calamities are likely to happen in the upcoming future?
b.What policies and procedure could help in planning and execution?
c.What are the needs of the affected population? What are reasons for the distress or movement?
d.What needs are most urgent? What additional resources are needed?
Big Data Analytics Big Data Analytics can help organizations in obtaining better answers to the above types of questions and in this process enable them to make sound real‐time decisions during and after the event as well as help them plan and prepare before the event (see Figure ). Descriptive analytics (that describes the situation) could be used for describing the current crisis state, identifying needs and key drivers as well as advocating policies. Prescriptive analytics (that prescribes solutions) can be utilized in alert and dispatch, prepositioning of supplies, routing,
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 5/9
supplier selection, scheduling, allocation, and capacity management. Predictive analytics (that predicts the future state) could be utilized for developing forecasts around societal needs, surge capacity needs in an emergency, supply planning, and financial needs. Four types of data streams that could be utilized to develop such models are social media data, SMS data, weather data, and enterprise data.
Social Media Data The availability of data from social media such as Twitter has opened up several opportunities to improve humanitarian emergency response. Descriptive analytics from the data feed during an emergency could help create the emergency crisis map in rapid time and inform about areas of acute needs as well as movement of distressed population. This could help with rapid response into areas that need the most help. Furthermore, such data feed could also be used to predict the future movement of the affected population as well as surges in demand for certain types of products or services. A detailed analysis of these data after the event could inform humanitarian operations about the quality of response during the disaster as well as better ways to prepare for future events of a similar type. This could be in terms of deciding where to stock inventory, when and how many supply vehicles should be dispatched and also make a case for funding needs with the donors. Simulation using social media data could provide solid underpinning for a request for increased funding. Analysis of information diffusion in the social network could present new insights on the speed and efficacy of messages relayed in the social network (Yoo et al. [ 30] ). Furthermore, analyzing the population movement data in any given region of interest could provide valuable input for ground operations related to supply planning, positioning, and vehicle routing. Finally, social media data is coming from the public directly and sometimes may contain random or useless information even during emergency. There is an opportunity to develop advanced filtering models so that social media data are leveraged in real‐time decision‐making.
SMS Data Big Data Analytics can also be adapted successfully for SMS‐based mobile communications. For example, a number of areas in the United States have started using cell phone SMS to text subscribers about warnings and alerts. Timely and accurate alerts can save lives particularly during emergencies. Predictive analytics models can be developed to determine when, where, and to whom these alerts should be broadcasted in order to maximize the efficacy of the alerts. The usage of mobile alerts is gaining momentum in the case of sustained humanitarian response as well. For example, frequent reporting of inventory at the warehouse for food and drugs can reduce shortages. Analytics on these data could provide more nuances on the demand patterns which in turn could be used to plan for the correct amount and location of supplies. Mobile phone alerts have also shown to improve antiretroviral treatment adherence in patients. In such situations, there is a great opportunity to analyze what kinds of alerts and what levels of granularity lead to the best response from the patient.
Weather Data Most regions have highly sophisticated systems to track weather patterns. This type of real‐time data is useful in improving the speed of response, so that the affected population can be alerted early and evacuations can be planned better. It also has a lot of information for designing humanitarian efforts for the future. For example, by analyzing the data related to the weather changes along with population movement, one could develop robust prescriptive models around how shelter capacity should be planned as well as how the affected population should
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 6/9
be routed to these locations. So, rather than trying to reach a shelter on their own, an affected person can be assigned a shelter and directed to go there. Prepositioning of inventory at the right locations based on weather data could improve response dramatically as reflected by the actions of firms such as Wal‐Mart and Home Depot that have made it a routine process after successful implementation during hurricane Katrina. Finally, the weather pattern data could be utilized to develop predictive models around the needs of the population in the medium to long term. For example, the drought cycles in certain regions of Africa follow a typical time pattern. A predictive model around the chances of famine in those regions could then inform the needs and funding requirements for food supplements.
Enterprise Data Most large humanitarian organizations such as UNICEF have information systems that collect a large amount of data about their operations. Analytics on such data can be useful to develop robust policies and guide the operational decisions well. For example, in systemic and emergent humanitarian needs, analyzing the demand and prepositioning inventory accordingly has shown to improve the operational performance. Furthermore, the analysis of long‐term data could provide guidelines for surge capacity needed under different environments as well as predict long‐term patterns for social needs across the globe due to changing demographics and socioeconomic conditions.
As the Big Data Analytics models and techniques develop further, there will be greater opportunities to leverage these data streams in more effective ways, particularly, given that the accuracy of data coming out of the different sources may not have the same level of fidelity in a humanitarian setting. While data are available in abundance in the developed world, there are still geographical areas around the globe where cell phone service is limited, leave alone social media data. In those situations, models with incomplete or missing data need to be developed. Also the presence of multiple decentralized organizations with varied degree of information technology competencies and objectives limits their ability to effectively synthesize the different data streams to coordinate decision‐ making.
Concluding Remarks Big data has enabled new opportunities in the value creation process including product design and innovation (Lee [ 16] ), manufacturing and supply chain (Feng and Shanthikumar [ 10] ), service operations (Cohen [ 3] ), and retailing (Fisher and Raman [ 11] ). It is also likely to impact sustainability (Corbett [ 4] ), agriculture (Devalkar et al. [ 7] ), and healthcare (Guha and Kumar [ 14] ). In our opinion, humanitarian organizations are also well positioned to benefit from this phenomenon. Operations Management researchers will have opportunity to study newer topics and develop robust models and insights that could guide humanitarian operations and make them more Responsive, Impactful, Sustained, and Efficient.
Acknowledgments The author wishes to thank Gemma Berenguer, Anand Bhatia, Mahyar Efthekar, and Jarrod Goentzel for their comments on an earlier version of this study.
References
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 7/9
1 Berenguer, G., A. V. Iyer, P. Yadav. 2016. Disentangling the efficiency drivers in country‐level global health programs: An empirical study. J. Oper. Manag. 45: 30–43.
2 Buchanan, L., J. C. Lee, S. Pechanha, K. K. R. Lai. 2017. Mexico City before and after the earthquake. New York Times, September 23, 2017.
3 Cohen, M. C. 2018. Big data and service operations. Prod. Oper. Manag. 27(9): 1709–1723. http://doi.org/10.1111/poms.12832.
4 Corbett, C. J. 2018. How sustainable is big data? Prod. Oper. Manag. 27(9): 1685–1695. http://doi.org/10.1111/poms.12837.
5 Deo, S., M. Sohoni. 2015. Optimal decentralization of early infant diagnosis of HIV in resource‐limited settings. Manuf. Serv. Oper. Manag. 17(2): 191–207.
6 Deo, S., S. Iravani, T. Jiang, K. Smilowitz, S. Samuelson. 2013. Improving health outcomes through capacity allocation in a community based chronic care model. Oper. Res. 61(6): 1277–1294.
7 Devalkar, S. K., S. Seshadri, C. Ghosh, A. Mathias. 2018. Data science applications in indian agriculture. Prod. Oper. Manag. 27(9): 1701–1708. http://doi.org/10.1111/poms.12834.
8 Dillow, C. 2017. The hidden costs of hurricanes, Fortune, September 22, 2017.
9 Eftekar, M., A. Masini, A. Robotis, L. Van Wassenhove. 2014. Vehicle procurement policy for humanitarian deevlopment programs. Prod. Oper. Manag. 23(6): 951–964.
10 Feng, Q., J. G. Shanthikumar. 2018. How research in production and operations management may evolve in the era of big data. Prod. Oper. Manag. 27(9): 1670–1684. http://doi.org/10.1111/poms.12836.
11 Fisher, M., A. Raman. 2018. Using data and big data in retailing. Prod. Oper. Manag. 27(9): 1665–1669. http://doi.org/10.1111/poms.12846.
12 Gallien, J., I. Rashkova, R. Atun, P. Yadav. 2017. National drug stockout risks and global fund disbusement process for procurement. Prod. Oper. Manag. 26(6): 997–1014.
13 Gralla, E., J. Goentzel, C. Fine. 2016. Problem formulation and solutions mechanisms: A behavioral study of humanitarian transportation planning. Prod. Oper. Manag. 25(1): 22–35.
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 8/9
14 Guha, S., S. Kumar. 2018. Emergence of big data research in operations management, information systems and helathcare: Past contributions and future roadmap. Prod. Oper. Manag. 27(9): 1724–1735. http://doi.org/10.1111/poms.12833.
15 Komrska, J., L. Kopczak, J. M. Swaminathan. 2013. When supply chains save lives. Supply Chain Manage. Rev. January–February: 42–49.
16 Lee, H. L. 2018. Big data and the innovation cycle. Prod. Oper. Manag. 27(9): 1642–1646. http://doi.org/10.1111/poms.12845.
17 Lorca, A., M. Celik, O. Ergun, P. Keskiniocak. 2017. An optimization based decision support tool for post‐ disaster debris operations. Prod. Oper. Manag. 26(6): 1076–1091.
18 Lu, X., L. Bengtsson, P. Holme. 2012. Predictability of population displacement after 2010 Haiti Earthquakes. Proc. Natl Acad. Sci. 109(29): 11576–11581.
19 McCafee, A., E. Brynjolfsson. 2012. Big data: The management revolution. Harvard Business Review, October 1–9, 2012.
20 McCoy, J., H. L. Lee. 2014. Using fairness models to improve equity in health delivery fleet management. Prod. Oper. Manag. 23(6): 965–977.
21 Meier, P. 2012. How UN used social media in response to typhoon Pablo. Available at http://www.irevolutions.org (accessed date December 12, 2012).
22 Natarajan, K., J. M. Swaminathan. 2014. Inventory management in humanitarian operations: Impact of amount, schedule, and uncertainty in funding. Manuf. Serv. Oper. Manag. 16(4): 595–603.
23 Natarajan, K., J. M. Swaminathan. 2017. Multi‐Treatment Inventory Allocation in Humanitarian Health Settings under Funding Constraints. Prod. Oper. Manag. 26(6): 1015–1034.
24 Rowling, M. 2016. Aid efficiency bargain could save $1 billion per year. Reuters, May 23, 2016.
25 So, A., J. M. Swaminathan. 2009. The nutrition articulation project: A supply chain analysis of ready‐to‐use therapeutic foods to the horn of Africa. UNICEF Technical Report.
26 Starr, M., L. Van Wassennhove. 2014. Introduction to the special issue on humanitarian operations and crisis management. Prod. Oper. Manag., 23(6), 925–937.
3/22/2020 Big Data Analytics for Rapid, Impactful, Sustained, and Efficient (RISE) Hu...: UC MegaSearch
eds.a.ebscohost.com/eds/detail/detail?vid=2&sid=2956bf6a-6493-4df7-9888-5624b87bfb48%40sessionmgr4007&bdata=JkF1dGhUeXBlPXNoaWImc… 9/9
27 Swaminathan, J. M. 2010. Case study: Getting food to disaster victims. Financial Times, October 13, 2010.
28 Swaminathan, J. M., W. Gilland, V. Mani, C. M. Vickery, A. So. 2012. UNICEF employs prepositioning strategy to improve treatment of severely malnourished children. Working paper, Kenan‐Flagler Business School, University of North Carolina, Chapel Hill.
29 Vanrooyen, M. 2013. Effective aid. Harvard International Review, September 30, 2013.
30 Yoo, E., W. Rand, M. Eftekhar, E. Rabinovich. 2016. Evaluating information diffusion speed and its determinants in social networks during humanitarian crisis. J. Oper. Manag. 45: 123–133.
PHOTO (COLOR): Big Data Analytics and Rapid, Impactful, Sustained, and Efficient Humanitarian Operations
~~~~~~~~
By Jayashankar M. Swaminathan
Copyright of Production & Operations Management is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
EBSCO Connect Privacy Policy A/B Testing Terms of Use Copyright Cookie Policy Contact Us
powered by EBSCOhost
© 2020 EBSCO Industries, Inc. All rights reserved.