The Monthly Briefing

Slow reopening 🔓

After the big shock produced by COVID-19 disease fast spread, and the measures to fight its impact, lockdown begins to be lifted: the reopening of economic and social activities starts.

Step by step. With restrictions. In phases. Asymmetrically among territories.
A reopening nonetheless.

From now on, and until the availability of an effective treatment that can be supplied to millions of people, we will have to live with the virus while trying to resume economic activity. And now technology can also be a huge aid in two main areas. Firstly, in the detection and contact tracing of infected individuals. And secondly, in the development of predictive models regarding the evolution of the epidemic and the adequacy of health resources, among others.

For this, however, we need to clarify doubts concerning the use of applications that comply with basic personal data protection laws, and, that crucially provide us with reliable data. This particular crisis is teaching us a number of things. For example, that the use of data has a huge potential to help in such a crisis, but also that we have a lot to improve in terms of the quality of this data.

Ricardo Baeza Yates, ACM & IEEE fellow and CTO at NTENT, writes about the importance of working with quality data in this article in Medium (in Spanish). According to Baeza-Yates, both the time dimension of the infection, the measurement biases and the fact that we only have part of the data add a lot of noise to the analysis and may give inconsistent results. "There is no point in having open data if it is not of good quality," says Baeza-Yates.

The importance of testing and tracing

Via Statista.

As we all know, COVID-19 has an enormous potential to spread. Just look at how many people a single person in Seoul could have infected during a night out. Tomás Pueyo, engineer and entrepreneur, explains how we can make progress in testing and contact tracing in this third article on the COVID-19. As Pueyo says, "only with enough testing, you can enter the Dance phase".

AI models have also been caught off guard by this pandemic

The fast changes occuring in human behaviour over the last months are directly affecting the correct performance of several Machine Learning models. Such models are trained on normal human behavior, and are no longer working as they should be when adjusting to behaviours that were not contemplated during the design phase. “Machine-learning models are designed to respond to changes. But most are also fragile; they perform badly when input data differs too much from the data they were trained on. It is a mistake to assume you can set up an AI system and walk away”, says Rajeev Sharma, global vice president at Pactera Edge, in this article of MIT Technology Review.

Many companies are now facing a comprehensive review of their machine learning models, as they need to adapt to these changes in customer behaviour. In this sense, the current situation highlights the importance of having a team dedicated to this task. "You need a data science team who can connect what's going on in the world to what's going on the algorithms," says Rael Cline, Nozzle's CEO.

However, Sharma also points to an opportunity: “A pandemic like this is a perfect trigger to build better machine-learning models,” he says.

COVID-19 tracing apps and data privacy

Screen details of the app launched by the Government of Spain. Vía Government of Spain.

In the last few weeks we have seen how many mobile data collection applications have been developed to help users to report symptoms of Covid-19 and track the outbreak. But not all these applications work in the same way, nor do they have the same implications for the protection of personal data. In this article, Juan Murillo, a member of the Data Strategy team at BBVA and an expert in the use of ‘big data’ for the greater good, answers some key questions to understand how these applications work, how effective they can be in controlling the epidemic, and whether they address users' privacy concerns.

In this article from MIT Technology Review they have suggested tracking all the apps that are tracking us. It sounds paradoxical, but it is very interesting to see what privacy guarantees each of these applications offers, as well as knowing what technology is being used to carry out this tracking and their degree of transparency.

Caroline Buckee, a top epidemiologist at Harvard’s T. H. Chan School of Public Health, has assembled a consortium of infectious-disease researchers to make the data accessible to policymakers -data that they did not yet have-. Buckee and other professionals share in this article their impressions about using and working with confidential data, such as geolocation.


of the


"Our goal should be to have correct and complete data. Not having reliable data is like driving a car with a lot of mud on the windshield".

Ricardo Baeza-Yates, ACM & IEEE fellow and CTO at NTENT.


At BBVA Data & Analytics we are helping in the co-organization of the KDD Workshop in Machine Learning for Finance. The goal of this workshop is to bring together researchers and practitioners to discuss the applications of ML and AI to the finance industry. Submissions are still open until tomorrow, May 20, 2020, and the workshop date is Aug 24, 2020.

Last year we were at KDD presenting a short article in the workshop on "Anomaly Detection in Finance". Unlike many other conferences, for us KDD really speaks in the language of "data science". In fact, it describes itself explicitly as targeting the “Data Science” community. For more info about teh conference, visit the website.

*Due to the current situation, KDD is monitoring the COVID outbreak closely and will decide whether the conference will be virtual or not. The latest update can be found here.

Figure shows the evolution of market shares for broad expenditure categories. The red dash indicates the announcement of the lockdown. Shares are expressed as a seven-day moving average.
Via BBVA Research.

BBVA Research has used big data technologies to examine the impact of COVID-19 on consumption in Spain. The analysis has led to a collaboration between researchers at BBVA Research, the University of Cambridge, Imperial College London, and the University of Edinburgh. The aim is to exploit high-frequency/high-resolution transaction data from BBVA, that includes both credit cards and point-of-sales terminals.

The article shows a slight shift in expenditure prior to the national lockdown, but thereafter an immediate, substantial and sustained reduction in spending. The study focuses on the first few weeks of confinement in Spain. In addition, the BBVA Research team is publishing weekly reports with the evolution of the main indicators. The last report published was on May 14th.

Artificial Intelligence can help overcome a crisis such as the one generated by the Covid19 in various ways and from different spheres of society. Projects related to pharmacology or medical care can be supported by these technologies, but also other approaches from the perspective of mobility or logistics, for example. However, it is also true that changes are requires and that there is still a long way to go before Artificial Intelligence technologies can develop their full potential. It is likely that AI will be a determining factor for forthcoming pandemics, but it seems that it will not be so much in the one we are currently living in.

Unprecedented time?

Maybe this tweet expresses quite well what many people are feeling right now in many parts of the world.

For any question or suggestion, you can also write to

You can enjoy much more content related to data science, innovation, new solutions of financial analytics and how we work in our website:

Let's talk about it. Join the conversation on Linkedin.

© 2020 All rights reserved. BBVA Data & Analytics. Avenida de Manoteras, 44, 28050. Madrid.