word count: 373 words
Poverty is a global problem that affects hundreds of millions of people worldwide. Many people cannot afford the common goods/services we need to survive: food, water, healthcare, etc. According to the World Bank, approximately 689 million people live in extreme poverty meaning that they live on less than $1.90 or less a day. The obvious way to solve this problem is to map and analyze poverty, so that proper funding and resources can be allocated to those nations. However, there are a few salient harms as well as the overall complex nature of using data science to map out/analyze poverty.
The harms that are most outright are the ones that have to do with poverty itself. Many die due to the lack of certain resources (water, food, shelter). Poverty is a mix of all the unfreedoms that Amartya Sen talks about, which makes it that much harder to solve. The journey to mapping out poverty may be done in certain regions of the world, but the true question is: Can these predictive models and mapping work in regions where poverty is not as apparent or outright? One of the many harms in data science itself is reducing the people we are assessing down to a number or a dot on a map. The methods used to produce these maps, although impressive, are still not as personable as the census surveys. Another harm may be that some aspects of these data collections may be intrusive and actually have the possibility of putting researchers in harms way. Data analysis has done a good job of making data collection less intrusive, but this can always be improved upon. Now we come to how these maps and predictive models are constructed, which is very exciting and promising. Satellite Imagery (CDR and RS data as well) is a geospatial data science that produce detailed images of poverty. Whether it be assessing the night-light usage in an area or analyzing the conditions of roads to infer poverty in that region, satellite imagery is nothing shy of amazing. Using the abilities of Satellite imagery, I feel that we can answer the question of: Can these predictive models and mapping work in regions where poverty is not as apparent or outright?