On 3-6 September 2020 the United Nations Statistics Division, along with UNCTAD, UN Global Pulse, Marine Traffic, and CCRi, hosted the UN’s first AIS Big Data Hackathon. The challenge featured 17 teams selected from a pool of 43 applicants from around the world, including government and international organizations, students and academics, and the private sector.
The hackathon was designed to highlight the capabilities of the UN Global Platform, which contains 3.6 terabytes of data from ship automatic identification systems (AIS), made freely available to members and collaborators of the UN Global Working Group on Big Data for Official Statistics. The AIS data include information on ship location, speed, and type, and this dataset has a wide variety of secondary uses. The UN Global Platform offers real-time AIS data and historical data from December 2018 onwards.
Participants were encouraged to focus on two primary themes: the economic and trade impacts of the coronavirus pandemic, and the environmental impacts of the shipping industry. With the help of nine AIS experts, the top three overall teams and three student-only teams were selected. You can read more about the ideas put forth by the winning teams below.
Top overall teams
Team members: Thomas Dewilde, Dennis Grebasch, Tomi Kallava, Alexandra Kapustina, Gennadiy Obraztsov & Farid Rakhmatullin. Contact: email@example.com
Team Blue Carbon, a cross-disciplinary team of Wärtsilä employees from four different countries, set out to create a global map of estimated shipping emissions, visualized in an easy-to-use dashboard. Using the EU Monitoring, Reporting, and Verification of CO2 emissions (MRV) dataset, the team was able to estimate emissions for merchant ships worldwide. The team then estimated the geographic distribution of emissions according to vessel locations and activity using a machine learning model. They created an interactive map which provides granular information on areas with high emissions, such as shipping lanes and busy ports. The map displays aggregated emissions concentrations on tiles at various zoom levels, enabling global and country-specific visualisations of shipping emissions for different time periods. The dashboard also allows users to estimate shipping emissions for a custom polygon, thus making it possible to remove the shipping-related “noise” from external estimates of regional emissions. You can read more about the project here.
Team members: Hussein Ashry, Rebeca Cáceres, Javier Díaz, Javier Sanchez Galán & Gabriel Fuentes. Contact: firstname.lastname@example.org
Team Data Caliber Group was made up of members from academia, industry, and government from three different countries (Panama, Egypt, and Norway). The team generated AIS-based transit indicators for both the Panama Canal and the Suez Canal, and showed that these indicators can signal the health of international trade using a reduced sample of shipping data. Moreover, the team established a framework for just-in-time arrival policies to reduce vessel emissions based on operational measures. The team’s AIS-based transit indicators can be used to measure the success of efficiency enhancement and emissions reduction policies implemented for the Panama and Suez canals.
Team members: Xiwen Bai, Liangqi Cheng, Yao Hou, Shiguan Liao & Dong Yang. Contact: email@example.com
Team Barents Sea was motivated by the observation that port-related statistics are traditionally published on a monthly or yearly basis by port authorities or shipping companies. These statistics are usually heterogeneous, delayed, and not easily accessible. The team worked to redefine three key port performance indicators using AIS and other data sources. First, they estimated the real-time throughput of global ports (i.e., the amount of cargo passing through the ports). Second, in terms of port efficiency, the team proposed a congestion indicator for monitoring real-time congestion levels at global ports. They then built a connectivity index at the port level, taking traffic characteristics and network factors of ports into consideration. The team’s throughput estimates proved to be 95% accurate when validated against real data. To illustrate port dynamics (i.e. the changing environment of ports, the status of international trade), the team developed a performance monitoring platform that you can learn more about here.
Top all-student teams
Team members: Takumi Kamata, Liwen Muring, Minato Nakashima, Hoshi Tagawa & Kodai Yasuda. Contact: firstname.lastname@example.org
Team DogCat used AIS data to analyze the impact of COVID-19 on society from an economic and social perspective. The team focused on several different topics, including the influence of COVID-19 on: offshore stockpiling, emissions around the cape of Good Hope, container shipping, and bulk carriers. The latter in particular yielded interesting results as presented here. The research focused on large vessels, namely capesize cargo vessels and very large ore carriers (VLOC). These vessels often carry coal and iron ore, which were chosen as the target commodities. By comparing periods before and during COVID-19, the team found no significant changes in iron ore trade. The trading behaviours of the largest export country and the largest import country, namely Australia and China, did not appear influenced by the pandemic. On the other hand, coal was more susceptible to the influence of COVID-19. The most noticeable origin-destination (OD) change was between Indonesia and India: the trade volume and trade times decreased greatly as a result of the pandemic.
Team Fraunhofer CML analyzed the regional impact of COVID-19 based on AIS and stock market data, with a focus on finding correlations between stock data from companies in Northern Germany and AIS data in the North and Baltic Sea before and during COVID-19. A key assumption was that observing many anchored and moored vessels reflects low port traffic, which in turn reflects a low flow of goods in that area. The team hypothesized that the low flow of goods leads to a decreased stock value of the companies in the investigated area. They used an interactive time-based heat map to analyse the data and visualized the stock data and the anchored and moored vessels using charts. The team found no obvious correlation between the anchored and moored vessels in 2019 relative to 2020 and the corresponding stock data, therefore further investigation into the data is needed.
Team members: Debora Doimo, Rodrigo Elias, Feliphe Galiza, Gustavo Gambera & Luiz Roma. Contact: email@example.com
Team Deep Water designed a monitoring tool with which the UN and other private or public institutions would be able to manage and analyze risks affecting four different domains: ships, crew, cargo, and health-related aspects. The team proposed the use of AIS data in combination with external data sources to create indicators which reflect these four different perspectives, such as the probabilities that a ship will: generate pollution, be associated with poor mental health among passengers or crew, experience weather-related cargo damage, and incubate a disease. The goal of the tool is to enable Deep Water and the UN to manage important aspects of daily life in vessels around the world and to predict future crises.
We celebrate the hard work of all teams, and the overall quality of submissions received. As the top-ranked team, Blue Carbon has been invited to send one team member to the UN World Data Forum in Bern, Switzerland in 2021. The top all-student team, Team DogCat, will have the opportunity to conduct a teleconference with a high-ranking UN official to present their results. More generally, we hope to highlight the contributions of all the top-ranked teams by encouraging them to present their results to stakeholders throughout the UN system, and by working with them to incorporate their ideas and code into the UN Global Platform.
For those interested in learning more, Team Blue Carbon and Team Data Caliber will give brief presentations of their projects at the UN World Data Forum’s Road to Bern event on Thursday, 15 October 2020 from 8:30 AM to 10:15 AM EST. The registration is available at https://unstats.un.org/bigdata/events/2020/road-to-bern/default.asp.