Unforgettable in 2020: Infection curves, COVID-19 dashboards

Back in January 2020 Lauren Gardner codirector of the Johns Hopkins Center for Systems Science and Engineering was tracking measles data when a couple of her PhD students began talking about a new coronavirus in ChinaMany of these students were Chinese and they were already tracking it closely In a few hours over one evening one of Gardners students built an original Covid19 dashboard from scratchBy the end of February the Johns Hopkins global dashboard was getting a billion hits a day By the summer it was 45 billion hits a day and more than a billion feature requests to access the underlying data2020 has been a breakout year for visualizations at the intersection of data science epidemiology and predictive models focusing the energies of large groups of data scientists coming together to build tools for population scale accessIn the US the first case of COVID19 was diagnosed on January 20 2020 By January 22 Johns Hopkins shared its Covid19 dashboard publicly loaded with the location and number of confirmed cases deaths and recoveries for all affected countriesAll data collected and displayed was made public initially through Google Sheets and later through a GitHub repository along with the feature layers of the dashboardGardner was named one of the 100 most influential people in the world by TIME for her work developing a free and open website that has become a global goto resourceIn parallel predictive models took off providing more than just 101 trackingThese plunged into infection projections hospital bed availability lockdown scenario planning mandates for easing restrictions and so on Many of these models have open sourced access to the documentation and the codeWhen the White House coronavirus task force nudged us in the direction of one such from the Institute of Health Metrics and Evaluation IHME at the University of Washington we never stopped checking in even after the White House briefings ended abruptlyTask force member Deborah Birx pointed to the scary infection curves to tell Americans to mask up Trump told Americans masks are optional and in the spaces in between all hell broke looseBy December 10 the virus was killing one American every 30 seconds and 150 cases were being reported in the same minute The country is home to the worlds highest coronavirus caseload and deaths More than 18 million infected since January and more than 328000 dead and theres no sign of easing offThe math was deadly from the time this exploded in China We dont need complicated math to comprehend its rampage and yet infection curves based on imagined futures are splashed on the worlds screens a roller coaster trajectory of utter chaosDo models play a critical roleAbsolutely models provide a visibility into possible impacts and allow the common man an ability to view this and do the required needful planning says an open source and Community Data Scientist This is indeed a new high for models seeing visibility he told IANSData scientists everywhere agree that data accuracy its collection and the maintenance of predictive models have been incredibly complex and problematic although evolvingNow that vaccines are arriving a lot of developers are taking a hard look at the sunset clause for the models they created to serve as digital public goods during a health care emergencyIts not just about the timecost of maintaining models thats on their mind its also the trafficNot all models enjoy the wild popularity that the Johns Hopkins tracker does and when public use dwindles engineers say its hard to just keep going Initially the sense of control they got over things we cant control fuelled their energy but eventually real world application won outAlso models have had a hard time in the public squareBriefly during Spring the reliability of one model versus another was front page news and very quickly the mood swung to the other extremeIn many countries editors at leading news publications were often wary of publishing model projections because they feared government pushback People stopped sharing predictive model links with their friends and relatives when they got messages saying things like it looks bad I dont want to see this stuffThe combination of lack of trust in the source data and pandemic fatigue wasnt just at the household level Even governments jabbed hard In fact in the US those who did not trust the IHME model bet on far gloomier infection curves from the University of Pennsylvania modelPolitics too got in the way big time The Centers For Disease Control and Prevention Americas premier public health agency was effectively sidelined by MarchWith zero consensus on viable streams of data state governors went all in local universities and developers for doityourself models New York Governor Andrew Cuomo who won an Emmy for his daily press briefings during the states worst phase of the outbreak described as maddening the range of estimates from predictive modelsIn America the cultural flashpoints turned out to be even more so which made predictive models that much more slipperyDespite all the skepticism around how models can be used effectively and governments reluctance to endorse these algorithmically mediated projections data scienceled use cases from the 2020 pandemic have already revved up action around the next big thing in digital public goods Community modelsKeep scrolling to read more newsCatch up on all the latest Crime National International and Hatke news here Also download the new midday Android and iOS apps to get latest updates MidDay is now on Telegram Click here to join our channel middayinfomedialtd and stay updated with the latest newsThis story has been sourced from a third party syndicated feed agencies Midday accepts no responsibility or liability for its dependability trustworthiness reliability and data of the text Midday managementmiddaycom reserves the sole right to alter delete or remove without notice the content in its absolute discretion for any reason whatsoever