Fighting Poverty using Maps

poverty mapping and poverty maps

How poverty maps help in deciphering socio-economic landscape

Poverty is a social phenomenon that has plagued humanity since the beginning of society. It isn’t simply a lack of sufficient income or resources but poverty in its most extreme form is a condition in which a certain section of society is unable to fulfill even the basic necessities of life. As a result, understanding poverty in its basic form and poverty mapping for various geographies has been a topic of research and development in all government and academic circles.

In India, poverty is widespread. In 2010, it was reported by the World Bank that 32.7% of the total Indian population falls below the international poverty line of US$ 1.25 per day while 68.7% live on less than US$ 2 per day. The main outcome of poverty is hunger. Hunger’s seriousness can be understood easily from the fact that every year, 5.8 million children die from hunger related-causes around the world as per the FAO Hunger Report 2008. Put another way, around 16,000 children die each day.

In terms of human lives, this is a huge catastrophe that needs to be culminated as soon as possible. To fight poverty on a global scale, it is extremely essential that we are able to define, identify and quantify poverty. The geographic distribution of the ‘below poverty line’ (BPL) demography can help us to understand the cause, resolution and scale of poverty. Without these locational insights, it is extremely difficult to tackle poverty.

As a result, numerous ways have been devised by governments and development sector organizations to gain spatial insights into the social and demographic welfare parameters. Their aim is to formulate policies and plans at the grassroots level as per the insights and analytics. Now, the most accurate and widely used statistical method to measure welfare-related insights is called ‘poverty mapping’.

What is poverty mapping?

According to the World Bank, poverty mapping (or, more formally, small-area estimation) technique developed by Elbers, Lanjouw, and Lanjouw (2003) combines the detailed information contained in household surveys along with the representation of census data to obtain reliable estimates of welfare at small-area level. Besides, this technique can also be used to estimate other indicators such as inequality and poverty rates in the given small areas. All these estimates can then be represented graphically on a map using a scale of colors, which has given rise to the colloquial term poverty maps. Thus, poverty map is a geographical data set, which provides a detailed description of the spatial distribution of poverty within a particular region. It is developed by combining a number of data collected through remote sensing data imagery.

GIS helps in mapping poverty at scale

With the recent advancement in remote sensing data and GIS technology, it is possible to develop accurate poverty maps for any geographical location with very less on-ground operations. Thus we can create poverty maps at scale. Through a combination of remote sensing and census data, various socio-economic metrics can be extracted and analyzed. These socio-economic metrics include factors like housing conditions, water distribution, population congestion and density, slum footprint, etc. These metrics can be then fed into various economic and statistical models developed by researchers and health organizations throughout the world.

Case Study: Measuring poverty using roof material type

At Attentive AI, we tested deep learning based technology to measure the scale of poverty in rural regions of India. For this, we used the type of roof/roof material as a proxy for household income. For e.g thatched roofs are used by families who are relatively poor. These houses cannot withstand extreme weather conditions and are also unhygienic. On the other hand, costlier metal roofs are relatively more permanent and preserve the interiors against weathering agents. Concrete or tile roofs are the most expensive and can only be afforded by relatively higher-income families.

Left: Ground survey image of thatched roof on left; Right: Thatched roofs extracted through satellite imagery

 Left: Ground survey image of metal roof on left; Right: Same metal roof extracted through satellite imagery

On a locality/village level, the roof material, when analyzed throughout a spatial belt or a region, shows the general trend or the overall degree of poverty in the area. In a locality, the ratio of the number of thatched and metal tiled roofs to the total houses in the locality can be used as an empirical poverty indicator. The greater this ratio, the poorer that belt or locality is and the more support it requires from government/development organizations to get a disease-free, weather-proof accommodation for the families living in the area. 

Using feature extraction on high-resolution satellite imagery, the Attentive AI team adeptly detected the thatched and metal-tiled roofs within a particular locality/village. Once these feature layers were extracted, the results were fed into a validated statistical model to assign poverty index to every locality/village.

The green marks indicate a thatched roof and the red marks indicate metal roofs. The left image contains a large number of thatched houses and the right side contains a good number of metal houses. This indicates that the region in the left image is poorer than the one on the right. Moreover, a poverty index can be computed based on the total number of houses and roof types to estimate a region-wide poverty condition. 

Advantages of poverty maps

Thus, poverty maps are an intuitive tool to study the socio-economic landscape of a region. Just like the analytic data layer which Attentive AI created over the satellite imagery for rural areas in India, such maps can be developed for any region on the globe. These poverty maps can then be combined with other macro-models to understand the economic situation comprehensively.

Once a region’s poverty map is ready, governments & development organizations can use it to analyze the socio-economic condition of the populace and define policies for the betterment of the state of affairs and help save human lives from the horrible jaws of poverty. 

Contact us to know more about poverty maps and how you can assess poverty and other demographic parameters using GIS.

Tags: , , , , , , , , , , ,

2 Responses

  1. Tabrez Akhtat says:

    Good observation .

Leave a Reply

Your email address will not be published. Required fields are marked *