Disaggregating economic inequality from space
(2019) In Student thesis series INES NGEM01 20191Dept of Physical Geography and Ecosystem Science
- Abstract
- Reducing inequality is number 10 of the United Nations sustainable development goals. Inequality in society affect economic growth, development and poverty reduction, and through these indirectly affect environmental degradation, food security and other major problems facing society today. Analyzing the temporal and spatial dynamics of economic inequality requires comparable data on finer scales. In most places, such detailed data is not available due to the high cost of production and uncertainty in the data and methods used to produce the inequality indices.
In order to fill the data gap, this thesis attempts to develop an inequality index from nigh time light (NTL) satellite images. Previous studies suggest the use of NTL images as... (More) - Reducing inequality is number 10 of the United Nations sustainable development goals. Inequality in society affect economic growth, development and poverty reduction, and through these indirectly affect environmental degradation, food security and other major problems facing society today. Analyzing the temporal and spatial dynamics of economic inequality requires comparable data on finer scales. In most places, such detailed data is not available due to the high cost of production and uncertainty in the data and methods used to produce the inequality indices.
In order to fill the data gap, this thesis attempts to develop an inequality index from nigh time light (NTL) satellite images. Previous studies suggest the use of NTL images as a proxy for missing data on SE-variables such as GDP and population. The advantages of these satellite-based proxies for measures human society suggests that similar methodology can be used to develop a proxy for inequality and thereby solve the data gap in this area.
The results presented in this paper confirms the use of NTL to model wealth, however modelling economic inequality is more complicated, and introduces problems based on the modifiable area unit problem and specific problems related to the definition of the Gini coefficient, a measure of inequality. These problems are reflected by the weak and in some cases non-significant associations found between the modelled and the actual Gini coefficients.
The results of this thesis show that both the correlations between actual and modelled Gini coefficients, the distribution of error as well as the mean absolute error varies depending on spatial scale of the analysis. Although spatial scale is suggested as a controlling factor, some patterns between the scales are still unexplained. Further investigation into other explanatory variables as well as avoiding errors due to discrepancies between different datasets might clear up the weak trends suggested in this thesis. (Less) - Popular Abstract
- It is vital that we work towards a more equal world in order to reach United Nations sustainable development goals. Inequality in society counteracts development, poverty reduction, economic growth and indirectly affects the affects society has on the environment. We need to understand the patterns, causes and effect of inequality in society in order to work towards a more sustainable future. However, understanding these patterns requires data with higher resolution in time and space. This data does in most cases not exist because it is expensive and labour intensive to produce. And even the existing data is difficult to work with because many different methods are being used to estimate this inequality.
To solve this problem this... (More) - It is vital that we work towards a more equal world in order to reach United Nations sustainable development goals. Inequality in society counteracts development, poverty reduction, economic growth and indirectly affects the affects society has on the environment. We need to understand the patterns, causes and effect of inequality in society in order to work towards a more sustainable future. However, understanding these patterns requires data with higher resolution in time and space. This data does in most cases not exist because it is expensive and labour intensive to produce. And even the existing data is difficult to work with because many different methods are being used to estimate this inequality.
To solve this problem this thesis suggests using satellite images as a proxy for traditional inequality data in order to produce a common inequality index that can be reproduced with as good a temporal and spatial resolution as the images have. Previous studies have shown that satellite images showing nighttime lights can be used as an estimate of socio-economic factors such as wealth and population. The light emission originates mostly from streetlights, industry and private homes, their distribution therefor represents the access/consumption of electricity, and thereby indirectly where people live and where people have money for electricity. Looking at the difference between electricity access in different areas could therefore be a good proxy for economic inequality in society.
In this paper I have confirmed that nighttime light images is a good replacement for traditional wealth statistics such as GDP. However, modelling an inequality indicator from the images is more complicated than wealth. The results of capturing economic inequality with the methodology show weak of nonexistent association with validation data. This is because this variable is even more sensitive to the unit area of the analysis than when modelling GDP. Furthermore, small differences in the inequality measures used in the analysis could be responsible for the weakness of the relationship.
In most cases my model underestimated inequality, especially in areas with high inequality. The patterns of error suggest that a driving factor in economic inequality was not capture by the light data. More research is needed in order to develop this methodology so that it might accurately capture inequality in society. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8997044
- author
- Maegaard Elvekjær, Neija LU
- supervisor
-
- Jonathan Seaquist LU
- Ola Hall LU
- organization
- course
- NGEM01 20191
- year
- 2019
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- Visible Infrared Imaging Radiometer Suite (VIIRS), night-time lights, Gross Domestic Product (GDP), economic inequality, regional inequality, spatial inequality, Gini coefficient, spatial scale
- publication/series
- Student thesis series INES
- report number
- 494
- language
- English
- id
- 8997044
- date added to LUP
- 2019-10-24 16:08:00
- date last changed
- 2019-10-24 16:09:37
@misc{8997044,
abstract = {{Reducing inequality is number 10 of the United Nations sustainable development goals. Inequality in society affect economic growth, development and poverty reduction, and through these indirectly affect environmental degradation, food security and other major problems facing society today. Analyzing the temporal and spatial dynamics of economic inequality requires comparable data on finer scales. In most places, such detailed data is not available due to the high cost of production and uncertainty in the data and methods used to produce the inequality indices.
In order to fill the data gap, this thesis attempts to develop an inequality index from nigh time light (NTL) satellite images. Previous studies suggest the use of NTL images as a proxy for missing data on SE-variables such as GDP and population. The advantages of these satellite-based proxies for measures human society suggests that similar methodology can be used to develop a proxy for inequality and thereby solve the data gap in this area.
The results presented in this paper confirms the use of NTL to model wealth, however modelling economic inequality is more complicated, and introduces problems based on the modifiable area unit problem and specific problems related to the definition of the Gini coefficient, a measure of inequality. These problems are reflected by the weak and in some cases non-significant associations found between the modelled and the actual Gini coefficients.
The results of this thesis show that both the correlations between actual and modelled Gini coefficients, the distribution of error as well as the mean absolute error varies depending on spatial scale of the analysis. Although spatial scale is suggested as a controlling factor, some patterns between the scales are still unexplained. Further investigation into other explanatory variables as well as avoiding errors due to discrepancies between different datasets might clear up the weak trends suggested in this thesis.}},
author = {{Maegaard Elvekjær, Neija}},
language = {{eng}},
note = {{Student Paper}},
series = {{Student thesis series INES}},
title = {{Disaggregating economic inequality from space}},
year = {{2019}},
}