Skip to main content

[Research Team Insight #2] Project MaPPPing and the role of Research Team - 5기 권경민, 6기 김예빈


1.      What is Project MaPPPing

 

SDP is constituted of students who have great interest in sustainable development, and we refer to the SDG goals articulated by the UN as the global standard for sustainable development. This year, we have specifically focused on goal number 9 and 11 that emphasizes the role of infrastructure in sustainable development [1]. We believe that sustainable and resilient infrastructure is at the heart of sustainable growth, so we have launched a project called Project MaPPPing in December 2019.

 

There are three phases to Project MaPPPing and we have been working on one phase at a time. During the winter of 2019-2020, SDP launched the project with Phase 1, and we have continued on with Phase 2 since the Spring semester of 2020. Both phases have been carried out as a collaboration with the World Bank Group PPI team. Working as a data reviewer and analyst supporting the World Bank Group, we have been updating data on PPI projects in middle- and low-income countries. Also, since the outbreak of COVID-19, we have been collecting data on infrastructure projects that have been disrupted by the pandemic.

 

The cooperation between SDP’s research team and tech team is a distinctive feature of Project MaPPPing since we aim not just to create a database on PPI projects, but also to create an AI tool that can automatically assess the status of PPI deals by looking at news articles.

 

2.      The Role of Research Team

 

The phase 2 can be separated into two parts: status tracking of PPI projects and tracking the delayed/canceled deals of PPI projects due to COVID-19. For better understanding, the role of research team will be described based on the latter since most of it overlaps between two parts.

 

At the initial stage of the project, research team and tech team gathered around to set the deadline and overall timeline. Given that the objective was collecting failed PPI deals in developing countries, each team specified their own schedule based on their work.

 

The main role of research team changed as the project proceeded. At the beginning, what we had done was preparing data to train the AI. With the aim to create AI tool to sort out the delayed/canceled deals due to COVID 19, we first arranged the keywords for crawling. The next process was giving weights to them respectively so that we could verify the necessary keywords. We also provided answer sheet for AI training by searching and accumulating some amount of urls with articles we are looking for. After receiving the first version of AI tool, the next work we’ve done was enhancing capacity. We classified frequently appearing websites but with meaningless contents to eliminate from the tool and also categorized keywords for minus weights.

 

The creation of relatively accurate AI tool, called as SDP tool, could be the point where the role of research team took a new turn. Our role was more of a research analyst. After studying the World Bank PPI Methodology, we reviewed the data in the tool and sorted out useful PPI deals to send it to the World Bank Group. This process continues until now, where phase 3 has been started, to keep updating and accumulating data.  

 

3.      The Learning Process

 

The process is not easy and there are hardships to doing this project. Mainly, we struggle finding relevant data on open source. This is especially true when we are searching for the current status of specific projects in specific countries. For example, we would search for the current status of a 400 MW LNG power plant in Thaketa, Myanmar, typing in keywords like the project name and the type of PPI in the search engines. Often, we come across search results that are irrelevant or there is just no data on that specific project. Sometimes, even if we did find articles with relevant information, we would not be able to fully access the site without subscribing and or paying.

 

Through this process we are continuously trained to be a good researcher. In order to find relevant information, we try out different combinations of keywords that gets us to the best possible search results. For example, we would type in the place name (city) with the infrastructure type (hydropower plant) instead of just putting in the project name. Also, we have found that typing in keywords in that country’s language rather than in English often gets us to more relevant information. In that case, we would utilize Google Translate to read articles written in different languages and try to get information on the project status.

 

What maximizes our learning process is the exchange of thoughts within the research team. During regular sessions, we share our experiences and insights that we obtained from researching. If someone found an effective combination of keywords, they would share it with the rest of the team. If someone had trouble researching specific projects, other teammates would suggest different ways of doing the research. If someone was unsure whether a project should be classified as operating or delayed, other teammates would look into the article together to determine its status. Likewise, the research team relies on each other and collaborates to solve problems and build skills.

 

In addition, while doing research for Project MaPPPing, we naturally become informed about recent developments in the field. This helps us improve our understanding of the general trend of infrastructure development in the developing countries along with specific details on certain projects. For example, as we are researching infrastructure projects that have been disrupted by COVID-19, we are aware of the various reasons ranging from the lack of workers and or supplies to the decrease in demand for electricity that deter the development of certain projects in certain regions. Having a broad knowledge on recent developments is important and helpful for us to grasp the dynamics of sustainable growth in real life.

 

 

4.      Project MaPPPing Phase 3

 

Project MaPPPing, which started in the end of 2019, has entered its final phase on September,  2020 along with the recruitment of the 6th members of SDP. With more people engaging in the project, the contents have been enriched and the process is speeding up to show its final work at the end of 2020.

 

The phase 3 has the title of “Visualization of the DB” and the visualization will take the form of failure map. It shows delayed or canceled hard infrastructure PPP (public private partnership) deals on the map so that various stakeholders can self-access to intuitive information on PPP infrastructure. PPP is considered as the most effective method of implementing and the key of sustainable development of infrastructure. Countries suffering from financial shortage, mostly developing countries, can partner up with corporates to distribute the financial burden of building power plants, dams, or roads essential for country development. Also, as the government can easily regulate PPP based infrastructures, those facilities can reinforce public purposes rather than solely pursing financial benefit. Through the steady and long process of phase 2, SDP’s own database on infrastructure has been enriched, prepared to be illustrated on the map. Among these data, the map utilizes PPP infrastructure deals worldwide, both developed and developing countries and further explains the specific reasons of failure. It provides filter service which the users can use and draw their own insights.

 

SDP believes the visualization of failure map to be a meaningful attempt, since, at the macro level, it is necessary work to contribute to sustainable development. As mentioned above, sustainable development builds on the presence and resilience of infrastructure. By arranging the causes of failure, the future stakeholders can harness the data and suggest their own solutions to prevent themselves from following the same failure path. The accumulated data will eventually give deeper understandings and lessons to infrastructure sector which will be the key to its resilience and improvement. SDP is delightful to mark the major milestone in this process and will continue to strive to seek the ultimate means for sustainable development.     

5기 권경민, 6기 김예빈

[1] UN Sustainable Development Goals https://sdgs.un.org/goals

Comments

Popular Posts

[Research Team Insight #1] The A to Z of SDP Research Team

[10월호, 2020년] 세계시민의식과 SDGs에 대해 묻다: 강동렬 UN SDSN Youth Korea 총괄 인터뷰 - 4기 유재희, 이소정

 "개인의 움직임은 과연 얼마나 중요할까?" 위의 질문에서부터 본 인터뷰는 시작된다. SDP는 현재 지속가능발전 글로벌 리더가 되고자 거시적인 시각에서 에너지와 인프라에 관심을 가지고 활동하고 있다. 우리는 다양한 외부 기관으로부터 지지를 받으며 유의미한 성과를 내고 있지만, 혹 멀리서 커다란 모습에만 집중한 나머지 개개인이 일상에서 지속가능함을 위해 노력하는 미시적인 움직임을 놓치고 있는 것은 아닐까. 세계시민의식을 가지고 살아간다는 것은 어떠한 의미인지, 우리가 생활에서 실천할 수 있는 방법은 무엇일지 고민하게 되었다. 강동렬님은 세계시민교육가이자 지속가능발전활동가로, UN SDSN Youth Initiative 한국 총괄로 계시며 SDP의 활동에도 언제나 많은 도움을 아끼지 않으신다. SDP는 지난 8월, 재생에너지와 시민의식을 주제로 동렬님을 모시고 공개 세미나를 추진하였으나 코로나19 상황 악화로 인해 잠정 연기한 일이 있다. 우리의 질문과 맞닿아 있는 세미나를 기다리며, 동렬님이 중요하게 생각하시는 가치와 활동은 무엇인지 SDP가 작은 인터뷰를 통해 들어보고자 하였다. 4기 유재희, 이소정

[3월호, 2021년] #32 딥 페이크(Deep Fake)와 디지털 포렌식(Digital Forensic) - 7기 진희주

  딥 페이크(Deep Fake)와 디지털 포렌식(Digital Forensic) 7기 진희주 아래 사진 중 무엇이 실제 오바마의 사진일까 ? 네 장 모두 딥 페이크 기술을 통해 인공지능이 만든 오바마의 사진이다 . 딥 페이크 (Deep Fake) 는 Deep Learning 과 Fake 의 합성어이다 . 딥 페이크는 AI 를 이용해 특정인의 얼굴이나 목소리를 합성한 사진 , 영상 , 음성 편집물 및 그 기술을 통칭하는 단어이다 . 딥 페이크 기술은 새로운 유형의 신경망 ( 머신러닝 알고리즘 ) 발전으로 가능해졌다 .   딥 페이크 기술을 활용하면 유명인의 얼굴을 대역에 합성해 손쉽게 영상을 만들거나 세상에서 사라진 멸종 위기 종이나 과거의 사람들을 생생하게 재현할 수도 있게 된다 . 여러 사람의 얼굴을 통해 가상의 얼굴을 제작하여 사용하거나 스타일 트랜스퍼 (Style transfer) [1] 를 활용해 겨울을 여름으로 , 또는 화창한 날을 비 오는 날로 묘사하는 등 이미지의 환경적 맥락을 바꿀 수 있는 등 딥페이크 기술의 영상 • 사진 산업 분야의 활용도와 가치는 무궁무진하다 .   하지만 딥 페이크는 엄청난 힘에 걸맞게 악용 가능성이 높아 그 위험성에 대한 논란이 최근 활발하게 이뤄지고 있다 . 대표적인 악용사례는 유명 연예인을 합성한 불법 포르노 , 전•현직 대통령의 얼굴을 합성한 가짜뉴스 영상 등으로 사회를 큰 혼란에 빠뜨릴 수 있음을 보여준다 . 딥 페이크 범죄의 대상은 유명인을 넘어 점차 일반인에게까지 확장되고 있다 . 헤어진 연인 , 또는 지인의 얼굴을 다른 나체 사진이나 성인물 동영상과 합성하여 만든 딥 페이크 제작물을 온라인 상에서 유통하거나 이를 통해 협박하는 등 딥 페이크 영상은 점차 심각한 사회적 문제로 대두되고 있다 . 2019 년 4 월 국내에서 큰 논란이 되었던 텔레그램 N 번방 사건에서도 딥페이크를 악용한 사진이 거래 , 유포된 정황이 확인되어 그 심각성이 전면으로 드러나게 되었다 ....

[11월호, 2020년] Countermeasures against urbanization: Air pollution in Seoul - 5기 함이수

Countermeasures against urbanization: Air pollution in Seoul 55% of the world population resides in urban areas as of 2018, according to a UN report. The overly populated urban population has given rise to problems that people were not aware of in the past. Especially, the impacts of urbanization on the environment are slow to be seen thus has been accumulating, invisible to the human eye. What has been done to address these problems? Furthermore, how should we approach these issues henceforward?