Nova > Faculty & Research > Knowledge Centers > Data Science

The Nova SBE Westmont Institute of Tourism and Hospitality (WiTH) partnered with Nova SBE’s Data Science Knowledge Center (DSKC) to develop the most comprehensive platform about tourism in Africa: the Nova SBE WiTH Africa.

The project aims to create and share tools and knowledge to study the evolution of tourism industry in Africa and support its sustainable development. It will provide an extensive set of data and translate it into actionable insights fundamental for the different stakeholders.

Considering the role of the Tourism Industry for the economic and social growth, the project ambition is to help improve African countries' competitiveness through insightful tourism data, knowledge, training and community engagement supported by the Sustainable Development Goals framework.

Partners: Nova SBE WiTH - Westmont Institute of Tourism and Hospitality
This is an ongoing project that started in 2021.

The social sector represents an important sector of the Portuguese economy. However, the few data sources available provide little and disconnected pieces of information, a low level of interaction and poor visualizations. In order to overcome this gap of information, the goal of the project is to create a new and integrated database for the social sector to be at the disposal of the Portuguese society, serving NGOs, foundations, impact-driven companies, researchers, investors, Government agencies, among other stakeholders.

Partners: Fundação ”la Caixa” and BPI (Social Equity Initiative)

Link: http://basededadossocial.pt

This is an ongoing project that started in 2019.

This project is the winner of the second edition of the Data for Change competition, and is being developed with Agência Natureza Portugal (ANP/WWF), in partnership with Docapesca – Portos e Lotas, S.A.
The project aims to use data analytics to shed light on the trends of the reported and sold fish caught before and during the COVID-19 pandemic, in Portugal, in all national ports. The goal is to understand how the fishing activity evolved during and after the pandemic. This exploratory project is also an experimentation on whether a predicting tool is feasible and useful for a better and environmentally friendly management of the fishing industry.

Partners: Agência Natureza Portugal (ANP/ WWF), Docapesca – Portos e Lotas S.A., Fundação ”la Caixa” and BPI (Social Equity Initiative)

This is an ongoing project that started in 2020.

We developed a system that assesses the risk of individuals becoming unemployed for more than 12 months, at different points of their unemployment journey. This could allow employment counselors to develop early interventions to prevent long-term unemployment. This work was co-created with experts in labor and social work policy, applying a user-centric approach that includes counselors themselves in the design process. As part of the research, we delivered a model that: i) was temporally cross validated; ii) can be dynamically applied at any point of the candidate’s journey within IEFP; iii) is explainable for each candidate; iv) was analyzed for bias against social groups; v) has better performance than the current model in production at the IEFP. In addition, we: i) investigated how the current model at IEFP was used by the counselors, and their attitudes and beliefs towards this AI model, by conducting a survey and focus groups; ii) developed tools to integrate our model in IEFP’s information system, in collaboration with the IEFP (e.g., a database infrastructure, a web service, and a dashboard); iii) designed and implemented a field intervention to investigate how different interfaces would impact the way counselors interact with the model.

The project is in deployment at national level at IEFP, and it is planned to be completed in the beginning of 2021.

Partner: Instituto de Emprego e Formação Profissional (IEFP), Faculdade de Engenharia da Universidade do Porto (FEUP) and Fundação para a Ciência e a Tecnologia (FCT)

This is an ongoing project that started in 2018.

In pandemics and other natural catastrophes, the management of population flow and stocks is key. We partnered with NOS to develop an information product that aims to help municipalities visualize and manage events that may pose higher risk of COVID-19 spread, using mobile network data.

(Grant: PT2020. AAC 15/SI/2020 - SISTEMA DE INCENTIVOS ATIVIDADES DE INVESTIGAÇÃO E DESENVOLVIMENTO E INVESTIMENTO EM INFRAESTRUTURAS DE ENSAIO E OTIMIZAÇÃO (UPSCALING) NO CONTEXTO DO COVID-19)

Partner: NOS, Câmara Municipal de Cascais, Câmara Municipal de Lisboa

This is an ongoing project that started in 2020.

The Electronical National Official Journal (DRE - Diário da República Eletrónico) is managed by the Portuguese Mint and Official Printing Office (INCM – Casa da Moeda) and allows the access to all Portuguese Legislation. However, and in order to clarify a specific question, citizens need often to look at different diplomas, that demand a technical and thorough interpretation.
The partnership between DSKC and INCM rises as an opportunity to enhance the experimentation in the Natural Language Processing (NLP) field, in Portuguese and in the law context, by developing a conversational tool to support citizens clarifying questions about their Retirement. Together with a team of students from the Master’s in Business Analytics, a proof of concept of the “Decoding Legislation” project will be created, intended to make the law more accessible and interpretable for citizens.

Partner: Portuguese Mint and Official Printing Office (INCM)

This is an ongoing project that started in 2020.

This project is the winner of the first edition of the Data for Change program and was developed in partnership with the Associação Protectora dos Diabéticos de Portugal (APDP).

Diabetic nephropathy is a critical condition that describes the gradual loss of kidney function, which can lead to irreversible damage to the kidneys. According to the Portuguese National Diabetes Observatory, one in three people with diabetes develops nephropathy.

Postponing or preventing this condition translates into an improvement in the quality and extension of the life expectancy of diabetic patients, also impacting the public health budget given that, as the disease progresses, the treatments become more complex and expensive, reaching dialysis or kidney transplantation. As it is a condition that evolves silently in the early stages, the first signs of diabetic nephropathy are critical to prevent or slow down the disease progression.

We analyzed historical transactional data, recorded during patient visits to APDP, which allow for a sociodemographic and clinical characterization of diabetics. As a result, the DSKC team developed a model to predict the likelihood of developing level three diabetic nephropathy over a three- and five-year time horizons. In addition to the probability, the tool also indicates the impact of each variable analyzed on the final result, allowing medical teams to act proactively, thus preventing the onset or progression of the disease.

Partners: APDP (Associação Protectora dos Diabéticos de Portugal), ”la Caixa” Foundation and BPI (Social Equity Initiative)

This is an ongoing project that started in 2019.

Use your data to inform decision making and better serve your beneficiaries and customers! “Data for Change” is a pioneer program in Portugal, which aims to identify Portuguese Social Impact Organizations that have a social/ environmental problem that can be solved through Data Science.

The projects are developed in partnership with the Nova SBE Data Science Knowledge Center team. The proposed challenges can be as varied as possible, covering areas such as Health, Employment, Environment and Education.

The project may aim, for example, to optimize the allocation of patients to doctors, the identification of factors that lead to school dropout, the identification of the risk of recidivism of prisoners, the optimization of emergency services, the increase of efficiency intervention strategies with homeless people, the increase in vaccination rates, the forecast of long-term unemployment, among many others.

The “Data for Change” program is part of the Social Equity Initiative, a partnership between Nova SBE, “la Caixa” Foundation and BPI

This project is an annual competition that started in 2018 and is currently on its 3rd edition.

Crowdfunding social projects is an innovative and inspiring way to foster cooperation and collaboration within our society. This project aimed to determine the key factors that influence the success of a crowdfunding campaign. Based on the analysis of more than 2.4 thousand campaigns already carried out on the platform, the DSKC team created a model that can be used at different points in time, during the campaign. In addition to informing the predicted probability of success, the tool also points out the main factors responsible for the forecast, in order to better inform the PPL team. In this way, PPL can more assertively provide feedback to fundraisers on how to best set up their campaigns, as well as proactively react to ongoing ones, in order to maximise the chance of raising the capital required for their projects.

Partner: PPL Crowdfunding

This project took place between 2019 and 2020.

Developing a predictive support system to understand which individuals are at risk of becoming long-term unemployed, in order to assist IEFP's Counsellors to determine how to personalize their professional development plans best. The tool allows Counsellors to understand which factors contribute to an individual's risk, giving them insight into the system's decision-making process, and granting the agency to override the tool in favor of an expert human judgment when appropriate. We are currently preparing a pilot study that will provide a critical evaluation of Counsellor’s opinions about the tool, to better understand the extent to which it facilitates their daily interactions with unemployed persons, and to support future research in unemployment prevention. This work was co-created with experts in labor policy and social work, including the Counsellors themselves in the design process.

Partner: IEFP, FEUP and FCT

This is an ongoing projcet that started in 2019.

In the last few years, Portugal has been witnessing rapid growth in tourism which is ill-equipped to face hard to quantify challenges such as tourist congestions, loss of city identity, degradation of patrimony, etc. The project proposes a data-driven approach to the tourism management in the Portuguese territory through a set of tools to inform Portuguese policy-makers about the state of the tourism industry, going beyond the traditional survey data and research and providing complete analyses of this sector using Big Data sources. With the insights provided by the project, tourism decision-makers will have crucial information to design policies that foster sustainable and friendly cities, both for tourists and dwellers.

Partners: Turismo de Portugal and NOS

As Tuscany is one of the most popular tourist destinations in the world, the region, its resources and inhabitants have been put under extreme pressure. The project aimed to use machine-learning algorithms to group individuals and, thus, help to better understand the region’s tourists and improve tourism management.

Partners: Toscana Promozione Turistica and Vodafone Italy

This project was part of the Data Science for Social Good Europe – Summer Fellowship 2018.

The purpose of the project is to use the data made available by the IEFP to build a modelled system whose objective is to empower the IEFP and its local offices to better identify individuals at high risk of long-term unemployment, and to provide data-driven insights which will allow for a more efficient allocation of IEFP’s resources to respond to the needs of unemployed individuals.

Partners: IEFP and EAPN

This project was part of the Data Science for Social Good Europe – Summer Fellowship 2018.

Financial institutions around the globe are looking upon the development of a data driven strategy. This strategy usually implies the restructure of their current ETL processes, teams and also the adoption of new technologies. The project proposed the implementation in Python of an existing algorithm along with several other new techniques to predict the credit default rate. For this purpose, we prepared a seminar for the data analytics department, where we used their data and the existing business problem to present the most common libraries and methods. The seminar was one of several efforts developed by the client to start two transition processes: 1) From licensed software (SAS) to Python, 2) From traditional econometric models to machine learning techniques.

Partner: Cofidis

The ultimate goal of the project was to increase the vaccination rate for measles, mumps, and rubella in Croatia. In order to achieve this, the team predicted the risk level of each child of not receiving the required two doses of the MMR vaccine and develop clusters of the above-mentioned children, according to similar features that would allow for customized vaccination-promotion policies.

Partners: Croatian Institute of Public Health (CIPH), Institute of Public Health of Split-Dalmatia County, Andrija Stampar School of Public Health and Croatian Society for School and University Medicine

This project was part of the Data Science for Social Good Europe – Summer Fellowship 2018.

Data-driven management is increasingly on the agenda of many Municipalities. The amount of data that Municipalities generate themselves, plus the data available from local partners such as Police, Transportation Companies, Telecom, among many others, offer a huge range of opportunities to develop new knowledge that supports public policy. CCDE is the data hub for the region, aiming to provide meaningful insights for local challenges, to experiment new approaches and solutions within the region and to foster knowledge dissemination through the integration of youth and a series of events.

Partner: Câmara Municipal de Cascais

The project’s goal was to create a machine learning model using data from incidents recorded, road characteristics, speed and flow traffic data and weather data in order to predict, within a given time window, the probability of an accident. This project allows for a better allocation of road inspectors to road sections to patrol and decrease their response time to incidents.

Partner: Rijkswaterstaat