COVID-19 Series: Machine Learning, AI and Automation in Current Healthcare Demands

CloudTechInfoCloudTechInfo
Keynote Presentation at Smart City Expo World Forum
info@cloudtechinfo.com 2020 Copyright © and Confidential
https://smartcitiesexpoworldforum.com/covid-19-protection/

CloudTechInfo
“CloudTechInfo promotes Digital Transformation for a sustainable world
through resource optimization driven by Enterprise Architecture practices”
CloudTechInfo
Machine Learning, AI and Automation
in Current Healthcare Demands
By
Anup Kale

About the Speaker
Data Scientist at
CloudTechInfo and MAppsAI
Industry Experience in ML,
AI, IoT, RPA
PhD in Machine Learning
More than 20 publications in
Machine Learning, AI and Data
Science

Agenda
7 important COVID-19 Numbers
COVID-19 from Data Science Perspective
• Visualizations
• Data
• Predictions
Healthcare Use Cases from AI, ML and Data Perspective
• Zonal Clustering
• Cognitive Automation
• Remote health Monitoring
• Robotics for dispensing and serving
Key takeaways

855000+ Confirmed Cases in the world
17700 Recovered
41652 Deaths
190+ Countries
25 Million Jobs
$1T or more GDP losses
17% more internet use
7 Important COVID-19 Numbers (By 1st April)

• Your sneeze may produce up to 40000 droplets, the farthest spread to
8 meters away
• A cough may product up to 1000-2000 droplets, up to 6 meters away
• Speaking can produce 500 droplets every minute
• I tend to shake hands average 3 to 4 times a day
• I tend to touch my nose possibly 6 to 8 times an hour
• Mostly 0.5 m Social distance at other place including fast-food
counter, Caffe etc
• Physical interactions for work related reasons (health, massage,
physio etc.)?
7 Numbers in Personal Space

COVID-19 Resources for Data Science Innovation
Subject
Knowledge
Visualization Data Prediction

COVID-19 Body of Knowledge
About the Disease
Global, National, State and Regional Statistics
Myths and Rumours: Examples of toilet tissues, drinking raw tea etc.
Test facilities and their capacities
General Pandemic Knowledge
Vaccine Related Efforts

• John Hopkins University
• World Health Organization
• Worldmeter
• esri Australia
• Visual capitalist
• Bing
• Google
• COVID Visualizer
Visualization Efforts by Global Community

Comparing Three Important Visualizations
John Hopkins
• Country-wise
Spread on Map
• Drill down to
state level for
USA, Au, China
etc.
WHO
• Country-wise
Spread on Map
• Combined stats
with other
diseases
Worldmeter
• Tabular data for
the global cases
• Distribution by
demographic,
age, sex etc.
• Plots on cases
fatalities etc.

Data Perspectives
Business
Objectives/Goals
Epidemic
Management
Data Structure Time
Dependency

Data Types
© Anup Kale

• Find a web-based Public data (Example: WHO)
• Government Releases (example: NSW Health).
• Qualitative or quantitative research data from Universities (Example: John Hopkins)
• Searching the public repositories including GitHub
• Real-time feeds from sources like WHO, dashboards, news channels etc.
• Publishing houses
• Demographic data can be through government departments
• Public health data can be requested by health authorities for research purposes due to its governance
restriction. Some historical datasets may be available to download and consume.
• Several APIs are available but require confirming their usefulness.
• Suburb level statistical information can be found through councils
• Google Cloud has released a free public dataset
• Healthmatch clinical trial dashboard
• Australian COVID-19 Hub by esri does provide data specific to Australian geography
Reliable Sources of Data

Proposed Data Journey for COVID-19 Use Cases
© Anup Kale

Rapid Data Acquisition Processes
Robotic Process Automation
Programmatic Approach
© Anup Kale © Anup Kale

Exponential and SIR (Susceptible - Infectious - Recovered ) Models
Predicting the Pandemic Spread
© Anup Kale

Potential Use Cases and Case Studies from ML
and AI Perspective
Predicting the Pandemic Spread
using Machine Learning
Managing Health Request Queues
(Enquiry and Registration) using
Software Robots
Spreading Awareness (facilities,
symptoms, healthcare information
etc) using Chatbots and AI
Reducing Administrative Burden
using Software Robots
Automated Delivery (especially for
nursing homes) using Robots
Remote Health Support using IoT
and machine Learning Services

Automation Tasks Involved
Admin Clerk Automation
Robotic Process Automation
Reception Automation
Through Voice and Chat Bots
Patient Query
Management
Through Voice and Chat Bots
User Interaction
Management
Through hand-held devices/Apps

• Pre Automation Challenges: 2 X 2 weeks staff equivalent repetitive
work every week causing to incur costs, inaccuracies and delay of 2
weeks to receive entire data/report causing delay in communication
with customers.
• Solution: Created a Automation Anywhere Software Robot for the
equivalent business process to:
• Save $50K/year
• Reduce the potential response time to customers from 2 weeks to 4 to 6
hours
• Improved Staff moral by avoiding the repetitive tasks
• Further potential to start web scrapping more data to create more business
potential which may further increase revenue
Admin Clerk Automation Case Study by CloudtechInfo

Remote Health Monitoring using IoT
• Monitoring Vital Signs for Cardio Vascular diseases
using network connected medical devices
• Use of proactive outreach to customers using
wearable sensing
• Telstra’s Virtual Health Monitoring: replaces face
to face meeting through video conferencing
• Activity Monitoring
• Blood Pressure and Heart Monitoring
© Anup Kale

• Image Classification
• Text Mining of Historical Data
• Post Peak Management
• Emergency Service Management
• Response Management
• Pharmacy Supply Prediction
Further Use Cases

Conclusio
n
Data the new oil is very important, but refining, validation and
security measures are essential
Multiple free trial tools are available for a quick and ad-hoc solutions
Low coding, no coding or coding templates can be followed for a rapid
development
Remoting, Virtualization and Automation are the need of time
Rumours, panic and cost cut measures spread faster than a disease we
need to refer genuine sources of information
Takeaways

Thank You
Any Questions?
Email : anup.kale@cloudtechinfo.com

COVID-19 Series: Machine Learning, AI and Automation in Current Healthcare Demands

More Related Content

What's hot (20)

Similar to COVID-19 Series: Machine Learning, AI and Automation in Current Healthcare Demands (20)

Recently uploaded (20)

COVID-19 Series: Machine Learning, AI and Automation in Current Healthcare Demands

Editor's Notes