Big Data In Addressing Poverty

This post is also available in: Deutsch (German)

The Coronavirus recession has worsened poverty, and by 2030, up to two-thirds of the global extreme poor may be living in fragile and conflict-ridden economies. We are a population of just 7 billion, yet we produce enough food for 10 billion people, and two minutes of sun is enough to power a year’s usage for humanity. But still, 39 out of 1,000 children will die before their fifth birthday because they’re poor while the richest 1% own 44% of the world’s wealth. Doesn’t this make you angry? Luckily, poverty doesn’t have to be passed down to younger generations as it can be cured. Hence, in this article, we discuss how Big Data Science can contribute to poverty eradication.

Implications of poverty

She walked into her little home devoid of decent furniture and materials for her to properly prepare and serve dinner to her family, to freshen up and rest nicely after a hard day’s work at the sweatshop.

He walked into his little home as he returned home from a hard day’s work at the construction site. His children’s loving embrace, as much as their skinny bodies could allow, would be the only sign of hope he could see the whole day.

These experiences may have been a dramatic representation of society’s perception of poverty but you get the point. It really sucks.

Although there’s been progress in reducing poverty since 1990, over 737 million people, 10% of the world’s population, still survive on less than $1.90 a day, the World Bank’s definition of extreme poverty.

And it’s no surprise that the Covid-19 recession has exacerbated the situation for the poor through job loss, remittance loss, rising prices and disruptions in essential services like healthcare and education.

The saddest part about poverty is that it affects millions of children around the world. Nearly 385 million children, who make up half of the extreme poor population, according to a joint study by the World Bank and UNICEF in 2016.

Due to poverty, 63 million children aged 6-11 years, 61 million teens aged 12-14 years and 139 million teens 15 years and above, are not in school. That is a staggering number of minors missing out on their education.

Another statistic to ruin our day is that an average of 39 out of 1,000 children will die before their fifth birthday. It doesn’t help that most of these under-five deaths are caused by diseases like malaria, diarrhoea and pneumonia, which are usually caused by malnutrition, contaminated water and poor sanitation. These issues are all preventable.

Thankfully, all hope is not lost for people living in this era as Big Data has the power to assist in addressing such issues.

Poverty tracking driven by Data Science

In Part 1 of our Big Data-driven SDG initiatives article series, we briefly mentioned the World Poverty Clock, which shows the real-time estimates of people living in extreme poverty and monitors each country’s progress towards meeting the SDG 1 target, which is to end poverty by 2030.

This means that I could have taken the extreme poor population estimate from this poverty tracker while writing that dramatic introduction. Let’s look at the number now. At this very second of writing, it is…

World Poverty Clock's number of extreme poor people.
According to this screenshot I took, this is the World Poverty Clock‘s global estimate of people living in extreme poverty at the time of writing. Hmmm… Seems to be less than the World Bank’s estimate but it’s still a lot of people in poverty.

The data set comes from a standardised global poverty database, which includes data from the World Bank, the International Monetary Fund, World Data Lab, International Income Distribution Data Set, World Economic Outlook, and governments worldwide.

The methodology, which has been peer reviewed in the academic community, makes use of Lorenz curves to depict populations’ income or spending distribution, compute the speed of poverty reduction in each country and region, and compare it to the average speed needed to achieve sustainable development goal 1.

It looks like a lot of work and resources are poured into building and maintaining the World Poverty Clock. While this poverty tracker seems to be worth the time and effort, there’s another poverty tracker in the hood called Robin Hood.

However, unlike its global real-time data equivalent, Robin Hood is said to be New York City’s largest poverty-fighting organisation, which provides support for food, housing, education, legal services, employment, etc to poor New Yorkers.

So, it just has its own Poverty Tracker as a tool to monitor the hardships of New Yorkers and provide the right support to the right people.

According to the Center on Poverty and Social Policy of Columbia University, a partner of Robin Hood, the Poverty Tracker is based on quarterly surveys of detailed information about income, material hardships and health of 2,300 households, a number which was increased to 4,000 in 2015, in New York City.

By understanding the data presented in the poverty trackers, poverty alleviation efforts can then be directed to the groups in need.

Poverty alleviation

Big Data sources for poverty alleviation efforts include mobile phone data, satellite imagery and biometric data. Ultimately, all of such data origins can be integrated together like in business systems for giving more complete images of reality.

Mobile phone data can fill the gaps of traditional census surveys as mobile phones are commonly used in developing countries. As Exhibit A, Ivory Coast’s mobile penetration rate is 83%.

So data collected from mobile phone activity can be used as proxies to identify time-series changes in socio-economic conditions (e.g. in form of anomaly or concept drift detection). This can be done in real-time by correlating call details with wealth patterns, for instance. Such data can also be included in the poverty tracker data set.

As seen in Part 2 of our Big Data-driven SDG initiatives article series, a government-backed online platform can also be developed to gather mobile app data about the poorest areas of the country. Through this platform, villagers can key in updates about their socio-economic conditions and needs so that help can be sent to them.

Satellite imagery, on the other hand, can be combined with machine learning and data analytics to identify geographically-specific poverty trends.

An example of using satellite data for identification is a pilot study in Sri Lanka to see if characteristics like building density, building height, built-up area, number of roofs, nature of roofs, vegetation lushness, nature of roads and the size of the administrative area can effectively indicate wealth.

As for biometric data, India is said to be well-known for using biometric data to ensure that aid is given to the right people and corruption is prevented through its national unique identity card program, Aadhaar.

By using biometric data to create unique identity cards for its citizens, the poor communities can effectively access aid and subsidies with instant identity verifications.

If mobile phone, satellite and biometric data can be leveraged to identify who needs help and send help to them, such data can be used to increase financial inclusion for people without formal credit using mobile banking.

Additionally, the fight against poverty would be incomplete without addressing unemployment. Although it’s not a guarantee that a job will pull someone out of poverty as some jobs have wages that don’t even cover the necessities, a job provides an opportunity to earn a living.

And if decent jobs can be created and directed to those who need them, the number of people living in poverty might go down. All that’s needed is mobile phone data to identify unemployment trends.

In 2015, the Journal of the Royal Society Interface published a paper about how one’s phone and social media usage says something about one’s employment status by studying what happened to mobile phone data when there was a mass layoff at an unnamed factory.

Using Big Data Science to empower the poor is a necessity

Leveraging Big Data Science to help the poor get out of poverty is a matter of giving them the right tools and empowering them to take charge of their livelihood and socio-economic situation.

It may not solve everything but it’s a great start. With Big Data, we can track the number of people living in poverty, who they are, where they are, what they need to survive, and so on. Maybe there’s still a lot more to explore in poverty eradication and socio-economic policy-making.

Now that we have the technology and Big Data’s helping hand to take all these measures, all we need now are the governments and corporations to act and work together in favour of these measures.

On that note, the importance of eradicating poverty can’t be stressed enough. If you still don’t get it, then ask yourself the following:

Earth has enough resources that we can erase poverty.
Do you really wanna live in a world where the poor get poorer?