Using data to connect dots in war on Covid-19

 data analysis
Many countries are now relying heavily on data analysis to manage and contain the current Covid-19 pandemic. FILE PHOTO \ NMG 

Data is increasingly becoming the most valuable asset in tackling numerous challenges across the globe.

Many countries are now relying heavily on data analysis to manage and contain the current Covid-19 pandemic sweeping across the world.

However some countries including Kenya, have yet to fully grasp the importance of Big Data despite its capability to help improve lives in a number of crucial domains

These countries ought to borrow methodologies such as those recently published by global data consortium Deep Knowledge Group on efficient means of extracting value from the health data released by the Ministry of Health (MoH) every day.

Despite their significance, data by themselves do not provide a complete picture. For instance, while data on a number of new infections, deaths, recoveries, gender and travel history are crucial, the government will have to dig deeper into these data points and establish the level of quarantine efficiency, emergency treatment readiness, monitoring and detection.


As Kenya's Covid-19 cases keep soaring, with only basic data about them being shared by the government and analysed by local data firms, experts believe more in-depth data can be generated from new infections, quarantine centres and self-reporting apps.

Absa Bank Kenya's chief data officer Hartnell Ndungi opines that although Big Data is a new field to both the public and private sector, now is the time to utilise it alongside proper healthcare management to help reduce new coronavirus infections.

"We are at a critical point where while the Health ministry is doing well to update Kenyans on the situation, the country needs more data points for statistical analysis. We require more data to get more insights about the disease," he told Digital Business.

Accurate data on how long it takes for patients to heal, their pre-existing health conditions, where they live, how they have been moving around, how they are responding to quarantine measures as well as real-time tracking on the progress of those tested.

MoH has been warning Kenyans via its bulk SMS platform that anyone can catch the virus and advising them to wash hands and avoid crowded areas. However, more information can be shared to keep citizens alert and informed.

"Because we are a tech-savvy nation, cloud computing should be used to gather data and compute insights and make them accessible to everyone for free. We don't need a platform that is centrally managed by health officials," remarks Mr Ndungi.

“Those in quarantine facilities can also share information about the sanitation conditions around them, while reporting on their progress on the platform. This could effectively reduce infections in the stations.”

To get a clear understanding of the virus, he adds, the Health ministry should create a cyber-secure open source cloud platform and give access to anyone to self-report on symptoms as well as allow researchers to publish insights.

Many quarantined people have complained of poor living conditions, expensive fee and going for several days without being tested making such areas dreaded stations.

Experts say a balance between data privacy and health risk aversion must be struck, allowing the government to use geo-location data of infected persons to shield the masses against the virus.

Nations like Singapore, Israel, Taiwan and South Korea are using data from citizens' mobile phones to perform "contact tracing" – a method that uses trajectory and geo-location data from mobile phones to detect and isolate individuals who were in close proximity to someone who has been tested positive.

The Kenya Data Protection Act of 2019 provides for access of data belonging to private citizens when there is a matter of public interest or when national security is threatened.

In the current situation, the government may want to pursue real-time accurate contact tracing and monitoring of quarantined persons who have the temptation to avoid social distancing.

Kenya can use location data to trace contacts of patients under isolation, but this must be done responsibly while protecting the privacy of those contacts.

The approach used in war on the pandemic is based on the same data analytics that digital marketers use to identify and target potential customers with ads through smartphone apps, through geo-targeting and geo-fencing.

"Anonymised and aggregated data from telcos is a key data source to understand mobility patterns of populations, improve decision making and scenario planning during the Covid-19 pandemic ," says Mr Timothy Oriedo, founder of local data firm Predictive Analytics Lab.

This data, he says, can be analysed in near real-time and provide an overview of mobility patterns, especially citizens who have been moving in and out of Nairobi, Mombasa, Kilifi and Kwale counties where cessation of movement is in force.

"Data on socio-economic changes caused by the pandemic should be available, such as those who need food relief or water as well as movements of urban populations to the countryside," he says.

Correlation of insights and comparison with biomedical data from other countries will also be important.

"We should be in a position to know the different strains of the virus, how long it stays alive on a surface and how it is mutating. This will tell us why those who previously recovered test positive again," observes Mr Ndungi.

The use of machine learning, a subset of Artificial Intelligence, whose roots have penetrated into Kenya's corporate scene, will be key to build regression models and algorithms to understand scenarios.

"This can be used to test hypotheses such as older people are easily killed by virus. We can also go further and use deep learning to actualize augmented technology and cameras to inform on people with Covi-19 symptoms without coming close to them."

A combination of these techniques can predict Kenya's coronavirus curve, giving way for the easing of lockdown measures.