Role of Data Science During COVID Times
                     ROLE OF DATA SCI
ENCE DURING COV
ID TIMES
HENRY HARVIN EDUCATION
WHAT IS DATA SCIENCE?
Defi nit ion One:
Data science proceeds to emerge as one of  the var ious encouraging and in-demand profession pathways 
for  exper ienced special ists.  Present ly ,  fl our ishing data special ists understand that  people need to 
develop beyond some universal  ta lents of  invest igat ing massive quant i t ies of  data,  data mining,  and 
programming ski l ls .  To reveal  valuable knowledge for  their  businesses,  data experts need to master the 
ent i re spectrum of  the data science development c i rc le  and maintain a level  of  adaptabi l i ty.
Defi ni t ion Two:
Data science course  is  an interdisc ipl inary course that  ut i l izes experimental  techniques,  methods,  
a lgor i thms,  and methods to obtain benefi t from data.  Data scient ists consol idate a var iety of  sk i l ls—
including stat ist ics,  network sc ience,  and industry exper ience—to interpret  data secured f rom the 
interconnect ion,  smartphones,  c l ients ,  sensors,  including other references.  Data science exposes 
courses and provides penetrat ions that  manufacturers can pract ice to obtain better  choices and 
generate more innovat ive products and services.  Data is  the bedrock of  var iat ion,  but  i ts  s ignifi cance 
or iginates f rom the knowledge data sc ient ists can discover f rom i t  and then act  s imultaneously.
The rapid spread and global impacts of COVID-
19 can make people feel helpless and scared as 
the novel coronavirus escalates and forces 
them to change many aspects of their everyday 
lives.
However, people can feel glimmers of hope in 
these uncertain times by understanding more 
about how data scientists are working hard to 
learn as much about COVID-19 as they can.
DATA SCIENCE CAN GIVE ACCURATE 
PICTURES OF CORONAVIRUS 
OUTCOMES:
Medical  professionals and others must get correct and up-to -date 
information about how the coronavirus s i tuat ion changes day by day.  
Several  organizat ions,  including Johns Hopkins Universi ty,  IBM,  and 
Tableau,  have re leased interact ive databases that  off er real - t ime views of  
what’s happening with the v i rus.
Many of these sources pul l  f rom data provided by trusted bodies such as 
the U.S Centers for  Disease Control  and Prevent ion (CDC) and the World 
Health Organizat ion (WHO).  They a lso inc lude direct l inks to those places 
so that people have quick,  easy access to rel iable information.
Using these databases can inform people of  the number of  confi rmed 
cases,  fatal i t ies and recoveries.  Then,  whether a  person is  on the front  
l ines of  the coronavirus fi ght or  a concerned c i t izen try ing to stay 
informed,  they can get a l l  or  most of  the information they need in  one 
place.
DATA SCIENTISTS DEVISE A SPEEDIER WAY 
TO HANDLE CONTACT TRACING:
Contact tracing is an effective way to slow COVID-19. It involves getting in touch with a 
person’s close contacts after that individual tests positive for the virus and telling them to 
self-isolate. Contact tracing is time-consuming, although it’s getting easier as more people 
take social distancing seriously.
Data scientists and medical experts teamed up at Oxford University to make contact 
tracing even more efficient. The experts working on the project asserted that 
mathematical models showed them how traditional methods of contact tracing used in 
public health are not fast enough to thoroughly slow the spread of COVID-19.
They created a mobile phone-based solution to eliminate the need for people to call the 
contacts manually. Instead, those parties get text messages confirming the need for self-
isolation. The researchers clarify that their approach would be most effective if it gets 
support from national leaders and is not an effort primarily spearheaded by independent 
app developers.No nations are using this method yet. Given the market penetration of 
mobile phones and the familiarity people have with receiving texts, however, it’s easy to 
see why this approach makes sense.
Many people with COVID-19 have only mild symptoms or none at all. 
Plus, the classic symptoms include a fever and a cough — two issues 
not restricted to the coronavirus. These things could make it easier 
for people to unknowingly spread the disease. But, developers 
EVERYONE created an app that uses data-sharing to help medical experts learn 
CAN PLAY A more about the virus.
PART IN It’s called the COVID Symptom Tracker and already has at least 
200,000 users. People can and should interact with the app even if 
HELPING they are asymptomatic or do not think their symptoms are COVID-
SCIENTISTS 19-related. The more researchers know about the coronavirus, the 
FIGHT THE better equipped they are to tackle it.
CORONAVIRU
S People interact with the app to do a short daily symptom check-in. 
They also give their age and zip code, plus disclose any preexisting 
conditions. That information helps scientists determine the groups 
that are most affected or in danger. The app does not take user data 
for commercial purposes, but it gives it to people who are working to 
stop the coronavirus, including some at health organizations.
DATA SCIENTISTS USE MACHINE LEARNING 
TO FIND POSSIBLE CURES FASTER:
Besides the race to restrict the COVID-19 The team taking this approach says this 
spread, scientists are working as quickly as method is less costly than traditional 
possible to uncover effective treatments. ones, too. Humans are still part of the 
Two graduates of the data science program process because they have to test the 
at Columbia University have turned to gene sequences identified as most 
machine learning to help. The typical promising by the machine learning 
process of antibody discovery during a lab algorithm. However, using this 
takes years. This approach, however, takes expedited method could be crucial in 
only every week to screen for therapeutic efficiently finding interventions that 
antibodies with a high likelihood of success. work for coronavirus patients.
DATA SCIENCE CAN HELP 
TRACK THE SPREAD:
Data science specialists have also concluded that graph 
databases are instrumental in showing them how 
COVID-19 spreads. Each plan database explains the 
connections between personalities, situations, or 
objects. Scientists refer to each of those entities as a 
node, and the connections between them are the 
“edges.” The results give a visual representation of the 
relationship between things, if any.
In the youth of the coronavirus outbreak, Chinese data 
scientists built a graph database tool called Epidemic 
Spread. It allowed people to type in identifying 
information associated with the journeys they took, 
such as a flight number or even a car’s license plate. 
The database would then tell those users whether 
anyone with a confirmed coronavirus case took those 
same trips and may have spread it to fellow 
passengers.
MAKING PROGRESS 
AGAINST COVID-19
Knowing as much about the 
coronavirus as possib le wi l l  
save l ives.  These are only some 
of  the fascinat ing ways that  
data sc ient ists  are using their  
sk i l ls  to help.
OPEN SOURCE DATA SCIENCE ON FIGHT 
COVID-19 
(CORONA VIRUS):
With the spread of  COVID-19 becoming an ever more assert ive 
force in our l ives,  the healthcare data science community has 
an opportunity to p lay an important  role in  the mit igat ion of  
th is  emerging pandemic.  Chronic le has g iven acknowledgment 
to the before-ment ioned compl icat ions vessel  drast ical ly  
change those most harmful  consequences of  the before-
ment ioned infect ions.  Many ci t ies have imposed socia l  
d istancing measures,  c los ing any place where large numbers of  
people gather,  and further measures can be taken to help 
iso late and protect  the most vulnerable among the populat ion.  
To do so,  we must fi rst  ident i fy  who is  at  greatest  r isk,  which 
mot ivated my team to create an open-sourced project ,  the 
COVID-19 vulnerabi l i ty  Index.  This  COVID-19 Dict ionary is  an 
open-source,  AI -based imminent design that  d ist inguishes 
personal i t ies that  do ant ic ipate to maintain a heightened 
vulnerabi l i ty  before cr i t ical  complexit ies of  COVID-19.
The COVID-19 Index is  meant to assist  hospita ls ,  federal  /  
state /  local  publ ic  health agencies,  and other healthcare 
organizat ions in their  work to spot ,  p lan for ,  answer,  and 
reduce the impact of  COVID-19 in their  communit ies.  In this  
post ,  we’ l l  be going over the high- level  detai ls  of  th is  open-
sourced project .
DETAILED 
DESCRIPTION OF 
DATA SELECTION:
STEP 1
Making a Labeled Data Set Data on COVID-19 hospitalizations do not yet 
exist. While data begins to emerge, we can look at the affected populations 
and events that serve as proxies for the real event. Given that the disease’s 
worst outcomes are concentrated on the elderly, we can focus on medicare 
billing data. Instead of predicting COVID-19 hospitalizations, we can instead 
predict proxy medical events, specifically hospitalizations due to respiratory 
infections. Examples include Pneumonia, Influenza, and Acute bronchitis. We 
identify these labels by parsing medical billing data and searching for 
specific ICD-10 codes that describe these types of events. All predictions are 
made on a specific day. From a particular day, we look back in time 15 
months for features. We exclude any events happening within three months 
of the prediction date, due to the lag in medical claims data reporting. Any 
diagnoses within the last year become the features we use in all of our 
models.
STEP 2 Models There are hosts of model considerations that need to be made with these 
kinds of projects. Ultimately, we wanted these models to balance being as 
effective as possible, and still accessible to healthcare data scientists as quickly 
as possible. One of the reasons for choosing the data that we used in because 
medicare claims data is widely available to healthcare data scientists. If your 
organization has access to additional data sources, you may observe performance 
increases by incorporating such information. Balancing those considerations led 
us to create 3 models based on the ease of adoption and model effectiveness. 
The first is a logistic regression model using a small number of features. At Closed 
Loop, we use the quality Python data science stack.
The motivation for a very simple model is that it can be ported to environments 
like R or SAS without having to read or write a line of python. At low alert rates, 
the model performs close to parity with the more sophisticated versions of the 
model. The aforementioned white paper has all of the weights for the limited 
feature set, so it can be ported over by hand. ROC graph comparing the 
performance of all three models. The next two models are both made using 
XGBoost. XGBoost consistently gives the best performance for making predictions 
on well-structured data, and given the right data transformation, medical billing 
data has that structure. The first XGBoost model is featured in our open-sourced 
package. A pickled version of the model exists, so you simply need to build a data 
transformation pipeline that will get your billing data into the format specified in 
the repo.
Add instruct ions or guidel ines here.  You 
can I f  you can bui ld  a funct ion that  wi l l  
parse your data for  a specifi c code,  then 
you can s imply i terate through a l l  of  the 
codes.  That’s  the reason we selected a 
l imited feature set  for  the open-sourced 
model .  I t ’s  very eff ect ive,  whi le st i l l  
requir ing only a reasonable level  of  l i ft  
f rom the data p ipel ine standpoint.  We’re 
a lso g iving healthcare organizat ions 
access to our model  within the platform. 
This  vers ion of  the model  uses fu l l  
d iagnosis  h istory,  p lus a large set  of  
engineered features.  Perceive,  the ROC 
s l ider determines that  the open-source 
report  becomes an approximately a l ike 
appearance as the report  inc luded in our 
programput in the amount of  t ime a l lotted 
for  th is .
ERROR 404: DATA NOT 
FOUND!
Have i t  in  the bag? Data (or  i ts  lack thereof)  can 
be the biggest  and most over looked chal lenge 
when i t  comes to the adopt ion of  data sc ience.  
Many organizat ions don’t  have the necessary data 
to perform data science.  Legacy pract ices,  
common examples of  which include – data 
captured through physical  forms,  unstructured 
data,  no scalable IT infrastructure in  p lace to 
process data,  and data stored in remote s i los,  are 
the pr imary reason that  some organizat ions are 
not  even aware that  the data they have is  of  no 
pract ical  use.  Pr ior i t iz ing data col lect ion and 
dig i t izat ion of  data from exist ing sources is  the 
f ront l ine solut ion to this  problem. However,  i t  is  
a lso important  for  companies to explore new data 
sources whi le  enhancing data access ibi l i ty  for  a l l  
key stakeholders.
WHAT IS BUSINESS INTELLIGENCE – BI?
Business intelligence (BI) refers to the procedural and technical support that handles, buildings, and interprets the data 
provided by a company’s actions. Business intelligence (BI) is a broad term that encompasses data mining, process 
analysis, performance benchmarking, and descriptive analytics. Business intelligence (BI) parses all the data generated by 
a business and presents easy-to-digest reports, performance measures, and trends that inform management decisions. 
Origins of Business intelligence (BI) The need for Business intelligence (BI) was derived from the concept that managers 
with inaccurate or incomplete information will tend, on average, to make worse decisions than if they had better 
information. Creators of financial models recognize this as “garbage in, garbage out.” Business intelligence (BI) attempts 
to solve this problem by analyzing current data that is ideally presented on a dashboard of quick metrics designed to 
support better decisions. Most companies can benefit from incorporating Business intelligence (BI) solutions; managers 
with inaccurate or incomplete information will tend, on average, to make worse decisions than if they had better 
information. The Growing Field To obtain helpfully, Business intelligence (BI) needs to attempt to improve the efficiency, 
opportunity, and significance of data. These requirements mean finding more ways to capture information that is not 
already being recorded, checking the information for errors, and structuring the information in a way that makes broad 
analysis possible. In practice, however, companies have data that is unstructured or in diverse formats that do not make 
for easy collection and analysis. Software firms thus provide business intelligence solutions to optimize the information 
gleaned from data. These are enterprise-level software administrations intended to join a company’s information including 
analytics. Although software solutions continue to evolve and are becoming increasingly sophisticated, there is still a need 
for data scientists to manage the trade-offs between speed and the depth of reporting. Some of the insights emerging 
from big data have companies scrambling to capture everything, but data analysts can usually filter out sources to find a 
selection of data points that can represent the health of a process or business area as an entire. This can reduce the 
necessity to capture and reformat everything for analysis, which saves analytical time and increases the reporting speed.
You might have come across this cliché – COVID-19 
has accelerated the shift to digital. With ongoing 
lockdowns and a projected recession, businesses are 
struggling to keep up with day-to-day operations and 
making tough decisions like layoffs, salary-cuts, and 
Capex rollbacks. While the present seems bleak and 
the future looks uncertain, businesses find 
themselves amidst an unprecedented crisis with only 
one thing certain: The future is digital. While some 
industries quickly adapted to remote work and digital 
INDIA: RETHINKING tools, others had to deal with multiple challenges to 
maintain business continuity. Business leaders have 
DIGITIZATION AND DATA been busy with the adaptation of new operating 
SCIENCE IN COVID-19 models, optimizing business processes, measuring RoI 
WORLD: of various spending, and gauging long-term business 
impact through data science and data-driven scenario 
simulations.
GOVERNMENTS AND HEALTHCARE 
PROVIDERS WORLDWIDE HAVE ADOPTED 
DATA SCIENCE IN MITIGATING THE IMPACT 
OF COVID-19:
This has been possible through the digital tracking of patients to monitor disease spread 
through epidemic forecast models to allocate healthcare resources through molecular modeling 
in drug and vaccine discovery and more. Access to quality data and data science experts to 
apply enhanced techniques is proving to be critical for faster recovery. With no travel and 
reduced meeting hours, it could be an apt time for CXOs to rethink their future in this changing 
business environment.
DATA AND DATA SCIENCE ARE TWO KEY INGREDIENTS OF 
ANY DIGITAL OPERATING MODEL:
While data science might seem l ike a luxury today 
amidst  th is  struggle for  surv ival ,  i t  could be a 
diff erent iat ing factor in decid ing winners of  
tomorrow.  With a few vis ionary companies al ready 
ahead on the curve,  a l l  organizat ions must plan 
their  d ig i tal  strategy to adapt to the post-COVID 
normal  before i t  looks us in  the eye.  In  the same 
context ,  th is  art ic le  d iscusses the potent ia l  
roadblocks in the adopt ion of  data science and 
poss ible ways to s idestep them. Culture 
Conundrums The storming of  the Bast i l le!  Yes,  th is  
is  where we get  to complain about corporate 
cul ture.
Gut-based decis ion-making,  mult i tudes of  excel  
reports ,  never-ending budget forecasts,  and out-of-
the-wor ld sales targets!  But  things could be better 
i f  a l l  decis ions were backed by data and facts  so 
that  everyone could see the underly ing rat ionale 
whi le  support ing and contr ibut ing to the decis ion-
making process.  Undoubtedly,  commitment from 
the top leadership is  v i ta l .  A  top-down mandate 
alone can’t  ensure the wide use of  data science for  
decis ion-making throughout the business.  A 
bottom-up adopt ion to embed data science into the 
way the organizat ion thinks,  decides,  and acts  are 
necessary for  good results .
COMING TO END:
The whole world is participating in a 
fight against this pandemic. The 
healthcare data science community can 
have a big impact on combating this 
disease. There have been many 
excellent efforts to use data 
visualization and Carlo simulations to 
help combat the spread of this 
pandemic. We feel our model addresses 
a complementary and important aspect 
of health policy, identifying those most 
at risk. By combining the efforts of these 
and many other excellent efforts in the 
healthcare technology space, we hope to 
mitigate the effects of this terrible 
disease. If reading this article has given 
you ideas for ways in which you’d like to 
contribute, we encourage you to be 
locked.
GET IN TOUCH
https: / /www.henryharvin.com/
[email protected]
+91 15266266