26 Mar 2020
3 MINS READ
December 30, 2019: A Toronto-based startup BlueDot, generated an alert to its clients about the possible outbreak of a deadly virus with Pneumonia like symptoms. BlueDot is a platform built around AI, machine learning and big data to track and visualize the outbreak and spread of infectious diseases.
This was at-least a month before the whole world was to take notice of the deadly disease, that later came to be known as COVID-19.
Not just that, BlueDot was also able to predict spread of the virus from Wuhan based upon airline ticketing data and it identified the cities that will get impacted in the early waves of the contagion.
In such testing times, data, analytics and AI is more relevant and important than ever before. There are tons of structured, unstructured data that is getting generated every-day in which lie insights with power to cause major impact. Let us have a look at different data-sets available that can be lead indicators of such forthcoming storms and give us a real-time view of prevalent situations.
IoT devices like smart watches: Scraping internet-connected devices like smart watches to detect symptoms like elevated temperatures, heartbeat, blood pressure, etc. in people can provide an indication of an underlying disease spread if symptoms start showing in statistically significant population size. Symptoms seen suddenly in a common set of population can be signaled out for further examination.
Unusual internet consumption: Quality of global internet connection can help glean economic and social insights. There are companies which monitor millions of internet-connected devices to gauge internet speed across the world. A sudden deterioration of internet network can mean extra stress on the internet line and it is a signal to an underlying surge of something abnormal. In China, in the industrial region, the internet consumption had gone down during January and February which corelated with low industrial production. The internet consumption has again started to normalize now, giving us the hope of a resurgence. Even unusual consumption on Netflix like spotting abnormal episodes of Netflix binge-watching can be an indicator of prevalent situation in a geography which might not be getting covered by international media.
Road traffic data: TomTom’s road traffic data for various Chinese and Italian cities was leveraged by many firms to understand how they are affected by quarantines and movement restrictions. Road traffic data can give an idea of multiple aspects like whether citizens are abiding with imposed lockdowns, panic is gripping the city, life is getting back to normal and whether people are preferring to use personal vehicles due to sense of insecurity in using public means of transport, etc..
Pollution data: Pollution trackers are another valuable source of information. Pollution levels are dropping across continents wherever there is an impact of COVID. For instance, the level of Nitrogen dioxide (NO2) emissions, regularly posted on Nasa’s website, is a proxy for pretty much any post-industrial human activity. NO2 is just about any emission coming out of cars, industries and a majority of our daily activities and, it’s changed quite a lot over the past month, with a big drop on the east coast of the USA, where most of the industries are located.
Geographic positioning systems (GPS): GPS is an important source to track the spread of 2019-nCoV. GPS data can provide significant contextual insights. For e.g. combining data from vessel transponders with satellite images, a company was able to understand how many oil tankers are in anchorage in China, unable to deliver their cargo – an intimation both of how well China’s ports are functioning amid the pandemic, and of how well industrial production is keeping up.
Social Media datasets: Mining social media and news websites on resources like www.healthmap.org, a website that tracks mentions of public health incidents around the world can provide valuable insights. We can also leverage live tweet and news articles datasets to build a similar dashboard like the website mentioned.
Many firms are extracting public data from more than 10,000 official and mass media sources in multiple languages, processing the text using natural language processing (NLP) and machine learning techniques, and summarizing the findings in a concise manner.
Analysing these data-sets for predicting spread of this contagion is important as it can give cities extra lead time to prepare for the attack by allocating limited resources more effectively like increasing the staffing levels of healthcare workers or ensuring provision of equipment needed to face the deadly disease like the masks and oxygen cylinders. The Chinese government took the extraordinary step of building an entirely new hospital in less than two weeks just to handle the 2019-nCoV patients, a few weeks lead time through data and analytics can translate into thousands of lives saved. Analysis of these datasets can also help governments understand compliance to their lockdowns and overall sentiment score of their populations. It is also relevant for organizations like WHO to get a true picture of what is happening to prevent and fight COVID across countries.
Can you think of any similar datasets which can be lead indicators for understanding and managing COVID effectively? Please share your thoughts for the same.
About the Author
Kunal Deshmukh is a Cloud Analytics Specialist at Hexaware Technologies with experience in helping enterprises adopt Cloud and AI/ML for data insights.
BI & Analytics
13 Nov 2020
07 Sep 2020
11 Jun 2020
28 May 2020
08 May 2020
24 Apr 2020
13 Apr 2020
06 Apr 2020
31 Mar 2020
23 Jun 2017
06 Aug 2015
13 Jul 2015
28 Oct 2014
17 Apr 2014
24 Mar 2014
22 Jan 2014
20 Dec 2013
01 Nov 2013
26 Sep 2013
03 Sep 2013
26 Aug 2013
29 Apr 2013
04 Mar 2013
21 Feb 2013
04 Feb 2013
03 Jan 2013
26 Nov 2010
19 Mar 2009
Digital Assurance
02 Jan 2012
17 Feb 2012
Infrastructure Mgmt. Services
02 Mar 2012
06 Feb 2013
Digital Assurance, Enterprise Solutions
14 Feb 2013
18 Feb 2013
27 Feb 2013
Others
01 Mar 2013
Enterprise Solutions
05 Mar 2013
18 Mar 2013
Digital Assurance, Enterprise Solutions, Others
22 Mar 2013
12 Apr 2013
26 Apr 2013
13 May 2013
11 Jun 2013
17 Jun 2013
25 Jun 2013
19 Aug 2013
27 Aug 2013
10 Sep 2013
19 Sep 2013
24 Sep 2013
30 Sep 2013
01 Oct 2013
03 Oct 2013
19 Nov 2013
Enterprise Solutions, Manufacturing and Consumer
28 Nov 2013
03 Dec 2013
03 Jan 2014
27 Jan 2014
31 Jan 2014
12 Feb 2014
13 Feb 2014
20 Mar 2014
11 Jun 2014
Manufacturing and Consumer
26 Jun 2014
30 Jun 2014
10 Jul 2014
15 Jul 2014
16 Jul 2014
18 Jul 2014
26 Aug 2015
28 Sep 2015
07 Oct 2015
26 Oct 2015
07 Mar 2016
22 Mar 2016
13 May 2016
23 May 2016
Application Transformation Mgmt.
11 Jul 2016
25 Aug 2016
03 Sep 2016
14 Sep 2016
15 Nov 2016
22 Nov 2016
25 Nov 2016
Business Process Services
25 Apr 2017
Banking and Financial Services
18 May 2017
30 May 2017
27 Jun 2017
18 Jul 2017
26 Oct 2017
Healthcare, Insurance
28 Nov 2017
11 Dec 2017
25 Jan 2018
21 Feb 2018
14 Mar 2018
( Mandatory field * )
The information you provide will be used in accordance with our terms ofPrivacy Policy
Please Check on "I Agree" to register for the blog.