Due to the Covid-19 outbreak, it is critical to assess the impact of the pandemic on our society. Governments and businesses are looking to determine the implications of Covid-19 and how the economy will recover from the pandemic.
In this episode of ExpertsConnect, Dr. Mehrnoosh Vahdat from IBM Data Science & AI Elite and Dr. Deepak Srinivasan from Rolls-Royce R² Data Labs provide an overview of the Regional Risk-Pulse Index: forecasting and simulation project as part of the Emergent Alliance. They show you a set of tools and dashboards that are launched for experts and the public and explain how they worked together to analyse a broad set of health, economic, and behavioural data to provide new insights and practical applications for the global Covid-19 response.
High-level overview on Emergent and use cases with examples [2:00] | Mehrnoosh explains that she and Deepak have been working as part of the emerging alliance project, which is a non for profit, organization, trying to address the challenges of COVID-19. She further discusses that when an emerging Alliance was established, many companies joined this initiative, for e.g., IBM, Rolls-Royce, Microsoft, Google, and many other companies to bring resources, data, and technology for coping with the challenges of COVID-19 through methods of data science.
What is the Emergence Alliance project about? [2:59] | She further shares that in the emergent alliance project the aim is to supply governments with insight that can be used for supporting their decision-making of COVID-19 countermeasures. She also mentioned that when the various collaborative entities got together and scoped use cases, they quickly understood that for a full understanding of the COVID-19 impact, they needed to not only look at the health implications but also, they needed to look into the behaviour and the sentiment of the population. In addition, they examined how COVID-19 impacted the economy.
Example use cases [3:40] | Mehrnoosh further elaborates that building on these three main workstreams (health, behaviour, and economy), the emergence alliance team has worked on many use cases. Examples of use cases include: (1) analyzing how health impacted the behaviour and the policies that the government imposed. (2) morbidity, the government measures, and how these impacted the sentiment of the population. (3) How all these factors impacted the economy.
Tools [4:27] | Mehrnoosh later asserts that in all of the use cases, pipelines are built-in IBM Cloud Pak for Data and IBM Cognos Analytics is used for developing the dashboards. In addition, she argues that because they were dealing with dynamic data sources, of which the COVID data is changing every day. She highlights that the collaborative team needed to build an end-to-end pipeline that could ingest the daily data and build and update the analysis pipeline, and improve and update the dashboards. She also stated that they built a number of knowledge management tools to inform the users about the analysis of their use cases, for example, bringing a chatbot or data repository, etc., to better inform the user about the results.
Behaviour Analysis [6:53] | Merhnoosh further explains that the team tried to understand how people change behaviour and their sentiment during the pandemic. They focused on the mobility wheel, the mobility estimator that predicts mobility for the next nine days, given a set of lockdown measures. They also looked into the airport restrictions and retrieving information on country-specific regulations. In addition, they developed dashboards to inform the user of this information. She also mentioned that they examined how the tourism industry has been impacted and how people change behaviour regarding tourism, also, how the sentiment of the population has been impacted by looking to the news, and media content. For instance, understanding how the topics that are discussed in the media have been changing over time or in social media, understanding how people are being affected and their mental health or the reactions towards the vaccine. She further elaborates that with all this information, they try to understand how and when we will economically recover from COVID-19 by building an analytical model of the economy. And looking into how different proxies can be translated to shocks in different industries and how these shocks in a specific industry have been propagated toward the whole industry.
Presentation of E3 and Travel Restrictions dashboard [10:57] | Deepak asserted that he thinks some of these tools could be very useful for governments, public servants, business leaders, and policymakers to understand the impact of COVID19 on the economy. He further describes the E3 - emergent economic engine, which is a tool that allows you to understand the impact of a shock i.e., a decrease on economic activity on the economic output of a country.
Emergent economic engine (E3) [12:02] | Deepak explained that because of COVID-19 and the ensuing health scenario, many governments or almost all governments across the world had to introduce non pharmaceutical interventions e.g., lockdowns, which severely curtailed people's activities, especially economic activities. For e.g., the hospitality industry has been severely impacted. He highlights that the E3 will provide a mechanism or will be critical as a tool to empower stakeholders to understand how the economic shock will depress the economic output. The tool is accessible via a web-browser e.g., Chrome and Firefox. Deepak further demonstrated a number of use cases with the E3, which can be viewed at segment [12:56].
Emergent economic engine Benefits [17:22] | Deepak summarizes the benefits of the tool as follows: it's quite useful if you're looking to find out the approximate forecasting of economic output contraction. WIth the E3 you can use it to determine what the worst-case economic output contraction could be.
Travel Restrictions Dashboard [18:08] | Deepak explains that the function of the tool to showcase the entry limitations for foreign travellers for a selected country along with accessible travel destinations for that specific country. He used Germany as a test example, using the tool you can find out from which countries could you enter Germany without any problem. In his example, one could see that they are allowed to enter to Germany from Australia, for example. And some of the countries, there might be some restrictions are indicated by yellow or Amber, which in essence indicates that there are some restrictions in place.
Travel Restrictions Dashboard Benefits [19:55] | Deepak summarizes the benefits as follows. One could find out how the travel restrictions evolve over time. He also argues that there is a common thread in that it unites both of these simulation products, the emergent economic engine and travel restrictions dashboard is that you've got notions of both temporal dynamics, that is how the lockdown affects the economy or the freedom of travel movement across time and space.
Cookie Cutter [20:34] | Mehrnoosh demonstrates the cookie cutter, which aims to present health and economic data from many resources and in various countries.
Questions and Answers [27:00] | This segment was led by Kadian Davis-Owusu and it includes the answers to the following questions.
“[…] When we address the users needs, we we need to get constantly their feedback to improve our tools and to, to their needs. And that's the way that we are processing is, is to launch these tools for public, so they can first access, and then hopefully, they come back to us with feedback.” [30:40] Mehrnoosh
“[…] So I think what the current COVID situation has shown us that there is increased awareness among the public for data. Just as a personal anecdote, I mean, my parents, family members who are not exactly that much into data science or software. I think they have an increased understanding of data. And you have politicians across the world, for example, that come on television […] explaining why R0 is so important, and why the exponential curve really matters. So you know, you've got phrases like flattening the curve, and you know, exponential curve, and so on. So I think even the non-mathematically inclined general public, if I may use the term has become aware of data and the importance of data in today's world.” [31:20] Deepak
"So I think it's absolutely important that a team of data scientists that look at data of such scope and magnitude should be broad, interdisciplinary, and be as diverse as possible. […] I mean diversity with respect to race, age, sex, ethnicity, training, backgrounds and experience. [34:19] Deepak
“[…] It's absolutely important to have really inquisitive mind and to be perfectly willing to take criticism, and to really encourage critical group thinking so that you just don't bend the data to tell the story that you want to tell. But you kind of have have a broad overall global view of the data.” [34:53] Deepak
“So, for validation when we apply data science networks, depending on the method, we have also validation techniques. For instance, when it's a prediction, then we try to see how the model is performing on a set of data that is not seen to the model. For instance, if it's happening in the future to see how closely it gets to the real outcome. Or, for instance, if it's clustering, you would see some coherence metrics, how are these clusters are very separated and are representing like reasonably, if they are matching our expectation." [35:38] Mehrnoosh
“[…] in addition to all these scientific methods, it's also important to get validation from the expert community […] . In most of these use cases, if we could find experts and independent researchers from universities, in similar fields, we would reach out to them, ask them to review our work and tell them to give us their expert opinion. And then we will validate or even improve further.” [36:24] Mehrnoosh
"And another validation that is very important is to validate with users. […], how they can be improved to to address the pain points of the users better, how we can improve them further. And for that, we took the steps to launch, for instance, these dashboards and the tools for reaching to the users better.” [36:58] Mehrnoosh
“[…] We have the users available, for instance, we are we are working with them with government specialists, or some local authorities. we involve them in design thinking workshops. So from the beginning, we design the use cases with the stakeholders to see […] who are the end-users, what are their pain points, what are the needs, and then go further with validating the data requirements and formulating the hypotheses. So through the I would say through design thinking workshop that's really something that can clear up what the users look for.” [38:52] Mehrnoosh
Get Involved with Emergent Alliance
Get in touch with Mehrnoosh: https://www.linkedin.com/in/mehrnoosh-vahdat/
Get in touch with Deepak: https://www.linkedin.com/in/deepak-srinivasan-b8b3815/
If you have been empowered by the information shared on this podcast, then please clap using the clap emoticon below. Please feel free to ask any questions or to share your comments using the comment emoticon.