Pabna University of Science and Technology

Department of Urban and Regional Planning

Report on
Determination of Level of Services (LOS) at Nimtola Morr Intersection,
Pabna, Bangladesh
Juliya Khatun, Tazkiwa Ahmed, Mohammed Riyad, Md. Nahid Sheikh Joy, Mithila Farzana Lima,
Mst.Sumaiya Yasmin
Department of Urban and Regional Planning, Pabna University of Science and Technology, Pabna-6600

Tanmoy Mazumder
Assistant Professor, Dept. of Urban and Regional Planning, Khulna University of Engineering & Technology

Submitted By
211702,211704,211716,211732,
211739,201741
Session: 2020-21
3 Year 2nd Semester
rd

Dept. of Urban and Regional Planning

Pabna University of Science and Technology

Submitted To
Tanmoy Mazumder
Assistant Professor
Dept. of Urban and Regional Planning
Khulna University of Engineering & Technology

Submission Date:
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Nighttime Light Derived Assessment of Regional Inequality of

Socioeconomic Development in Bangladesh

Highlights

1. Determination of Regional Disparity through NTL data and Socio-Economic Indicator.

2. Finding a strong connection between NTL glow and city life, industries, and well-built
homes.
3. Sorting districts into low, medium, high, and very high gap levels using Coefficient of
Variation.
4. Showing NTL as an affordable way to track social-economic differences and guide smart
policies.
5. Uncovering a clear divide between urban and rural growth in Bangladesh.

Abstract
This study investigates regional disparities in socioeconomic development across Bangladesh
using Nighttime Light (NTL) data as a proxy for economic activity and urbanization.
Leveraging satellite-derived NTL imagery from Google Earth Engine (2018–2022) and
socioeconomic indicators from the Bangladesh Bureau of Statistics, the research maps
variations in development across districts and examines correlations between NTL intensity
and factors such as population, literacy, housing quality, and industrial activity. Correlation
analysis reveals strong positive relationships between NTL and urban population (r = 0.936),
industrial population (r = 0.944), and pucca housing (r = 0.912), while rural and agricultural
regions exhibit weaker associations. The Sopher Index highlights significant literacy-illiteracy
disparities, with districts like Dhaka and Chittagong showing exceptionally high inequality (SI:
0.9–1.54). The Coefficient of Variation further classifies districts into low (22), moderate (39),
high (1), and very high (2) disparity categories, underscoring uneven development. These
findings emphasize NTL’s potential as a cost-effective tool for assessing socioeconomic
progress and identifying regional inequalities. The study advocates for targeted policy
interventions to address urban-rural divides and promote equitable development in Bangladesh.

Keyword: Socio-Economic Indicator, Inequality Mapping, Sopher Index, Coefficient of

Variation, Correlation Analysis

1|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

1. Introduction
Regional inequality is widely regarded as a disadvantage of economic development. Accurate
assessment of poverty, especially at the regional level, is crucial for governments and
policymakers. Satellite imagery provides efficient and accurate spatial data at a low cost. One
example is nighttime light (NTL) data. NTL data can accurately estimate electricity
consumption, population, and GDP due to their strong correlation with human activities.
(Sangkasem & Puttanapong, 2022).Data from national censuses and surveys monitor
socioeconomic indicators, but in Bangladesh, it’s often broad and outdated. The census, last
held in 2011, updates every decade, while frequent surveys stop at district-level detail.
Breaking them down to subdistricts leads to unreliable estimates with high errors. Satellite-
based Night-Time Light (NTL) data emerges as a practical substitute, reflecting various
socioeconomic trends effectively (Wahed et al., 2020).

NTL data have been regarded as an alternative that can be used in the absence of economic
census variables (Lazar, 2011). (Jean et al., 2016) shows that using NTL data alone is not good
enough to differentiate between poor and ultra-poor regions since the low light levels in these
regions cannot be distinguished from noise in the data. Combining the NTL data with the
population data (Zeng & Zhang, 2019) categorize cities in the Sichuan Province of China with
respect to the level of urban development in those cities. The authors conduct regression
analysis to identify three “ghost cities” in the Sichuan province where insufficient urban
planning and imbalanced development has led to high housing vacancy rates.(Rahman et al.,
2019) categorize the cities of Bangladesh into 6 different clusters based on five spatial features
including the total NTL intensity. We observe the correlation between NTL and various
traditional indicators that are used to describe the demographic and economic conditions of an
area (Wahed et al., 2020)

Bangladesh faces stark regional differences in literacy, employment, infrastructure, and other
socioeconomic variables like income and education access, yet such disparities remain
underexplored. Night-Time Light (NTL) data provides a new way to study these patterns. This
study aims to map these variations across districts and analyze how NTL intensity correlates
with key socioeconomic factors. With limited research on this in Bangladesh, this work fills a
crucial gap. The goal is to highlight inequalities and test NTL as a tool to measure
socioeconomic progress effectively. Policymakers can use these insights to target development
efforts, while globally, it advances methods to measure progress using NTL.

2|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

2. Materials and Methodology

2.1 Study Area

Bangladesh, a rapidly developing South Asian nation, faces significant regional disparities in
socioeconomic conditions, urbanization, and access to essential services spanning latitudes
20°34'N to 26°38'N and longitudes 88°01'E to 92°41'E. The country’s development trajectory
is marked by pronounced spatial inequalities between urban and rural areas, driven by poverty,
unequal resource distribution, and unplanned urban expansion. By assessing spatial clustering
of economic activities and variations in NTL intensity, the study will highlight urban-rural
divides and regional inequalities.

Figure 1.Study Area Map (Source: Author, 2025)

3|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Figure 2.Study Area Map Indicating Lit Area (Source: Author, 2025)

2.2 Data Collection

The Night-Time Light (NTL) data, which represents artificial light emission at night and is a
measure of urbanization and economic activity, was derived for this study using Google Earth
Engine (GEE). We have taken relevant data indicating socio economic factor from the
Bangladesh Bureau of Statistics (BBS), linking variables (% lit area, mean values etc.) with
the spatial data. The variable’s data covers the period from 2018 to 2022.

4|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Table 1.Description of Variables with Data Sources

Variable Description Sources

1 Total Population Entire number of people living within a
specific area.
2 Rural Population Number of people live in designated
rural areas within a specific region.
3 Urban population Number of people living within the
defined urban boundaries of a region.
4 Total slum & The combined number of people
floating Population residing in both slum settlements and
those considered as part of the
"floating population" within a specific
area.
5 Financial Account Record of financial transactions in a
area
6 Dependency Ratio Average number of economically
(Total) dependent population per 100
economically productive population,
for a given country at a specific point
in time.
7 Population growth Indicating percentage of population
Rate increasing

8 Literacy Rate The percentage of the population aged

15 and above who can read and write (BBS, 2022)
with understanding.
9 Main Source of Include hydroelectric power,
Electricity thermoelectric power wind power,
solar power, and nuclear power.
10 Internet user Total Number of people use internet.
11 Semi-pacca House Amount of Semi-Pacca House of an
area.
12 Jhupri House Total number of house made of wood.

13 Pucca House Made of durable materials like brick,

concrete, cement, and steel
14 Kancha House low-quality, low-durability structures
typically found in rural areas.

15 Foreign Remittance Funds sent by migrant workers back to

their home countries
16 Population- Number of people who depend on
Agriculture agriculture, hunting, fishing, and
forestry for their livelihood.
17 Population- Industry Refers to the relationship between the
size and characteristics of a population

5|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

in a particular region and the types of

industries that are prevalent there,
18 Population-Service The total number of people residing
within a specific geographical region
that can be served by a particular
service.
19 Mobile-Banking Service provided by banks and
financial institutions that enables
customers to conduct financial
transactions remotely using their
mobile devices.
20 Own dwelling Unit Single, self-contained residential space
within a particular geographic location
that is owned by the person or family
occupying it. (BBS, 2022)
21 Not Having Own Do not own a house, apartment, or any
Dwelling Unit other independent living space within a
particular location
22 Mobile user Total Mobile phone user

24 Illiteracy Rate The percentage of people in a area who

are unable to read or write
25 Mean-of-intensity Mean or average value of intensity of Through ArcGIS
NTL image 10.8 from NTL
image
26 Max-of-intensity Maximum value of intensity of NTL Through ArcGIS
image 10.8 from NTL
image
27 Sum-of-Intensity Summation value of intensity of NTL Through ArcGIS
image 10.8 from NTL
image
28 % Lit Area Total lit area in percentage Through ArcGIS
10.8 from NTL
image

2.3 Data Processing

In this study, we utilized Google Earth Engine (GEE) to derive Night-Time Light (NTL) data,
which reflects artificial light emission at night and serves as an indicator of urbanization and
economic activities. The collected NTL data was transformed into the BUTM coordinate
system to better fit the study area. We also acquired the district boundary map of Bangladesh
using ArcGIS 10.8, ensuring broader spatial coverage for the analysis. Data processing in
ArcGIS was performed using three primary tools: Reclassify, Raster Calculator, and Zonal
Statistics. The Reclassify tool was used to replace NTL intensity values with class codes based
on set thresholds. The Raster Calculator tool manipulated the reclassified data to create

6|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

representative outputs. The Zonal Statistics tool then provided statistical indicators, such as lit
area percentage, mean values, and total light intensity within each district boundary. To
incorporate socio-economic factors, we attached relevant data from the Bangladesh Bureau of
Statistics (BBS), linking these variables with the spatial data. Finally, SPSS was used to
conduct a correlation analysis to determine potential relationships between NTL intensity and
the socio-economic factors, helping to identify key variables that influence urbanization and
economic activities.

2.4 Correlation Analysis

Correlation analysis helps determine how strongly night-time light intensity relates to socio-
economic factors, making it useful for assessing development, urbanization, and economic
activity. Since reliable ground data is often unavailable, night-time light serves as an alternative
indicator. A strong correlation suggests that higher values of a variable, like urban population
or industry, lead to brighter areas. This helps identify patterns in economic growth,
infrastructure, and disparities between regions. Correlation is used because it is simple,
interpretable, and effective in highlighting significant associations, making night-time light a
potential proxy for development trends.

Table 2. Range of Correlation Coefficient Values and the Corresponding Levels of Correlation.

Coefficient Values Correlation Coefficient values Correlation

0.80 to 1.00 Very Strong Positive -1.00 to -0.80 Very Strong
Negative
0.60 to 0.79 Strong Positive -0.79 to -0.60 Strong Negative
0.40 to 0.59 Moderate Positive -0.59 to -0.40 Moderate Negative
0.20 to 0.39 Weak Positive -0.39 to -0.20 Weak Negative
0.01 to 0.19 Very Weak Positive -0.19 to -0.01 Very Weak Negative

2.5 Sopher Index

Sopher's Disparity Index is a tool for measuring disparities between two groups or regions. It
uses a logarithmic scale to reduce the impact of base values, highlighting relative differences
over absolute ones. A value of zero represents complete equality, whereas higher values
indicate increasing levels of inequality. The index's figures have no standalone meaning and
are best understood when compared over time periods such as years or decades. Unlike
methods that examine variation within a single dataset, this index focusses on the differences
between two distinct data series. The formula appears below

7|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

x2 (200−x1)
DI= 𝑙𝑜𝑔 x1 + 𝑙𝑜𝑔 (200−x2 )

Where,
x1 = %Literate people in a city
x2 = %Illiterate people in a city
Sopher’s Disparity Index provides a logarithmic measure to evaluate inequality between two
groups or regions, with values indicating the extent of disparity. The following table outline
the theoretical ranges of the index and their corresponding levels of inequality, offering
framework for interpreting disparities across various types.

Table 3 Sopher Disparity Index and their Classification Based on

Disparity

Index Value Range Level of Disparity

0.41-0.6 Moderate Disparity

0.6-0.7 High Disparity

0.7-0.8 Very High Disparity

0.8-0.9 Extremely Disparity

0.9-1.54 Exceptionally High Disparity

2.6 Co-efficient of Variation

It is simple to grasp and contains all of the values in the dataset. While weighted data can
protect against outliers, it lacks a baseline for acceptable inequality. The coefficient of variation
is the distribution's standard deviation divided by

mean. CV ranges between 0 and ∞. A value of ∞ implies excessive inequality while a value of
zero suggests complete equality. This study employs the unweighted coefficient of variation
(CV) approach to assess inequities in socio-economy across regions. (Wasim & Munir, 2017)
uses this method to find inequality between and within regions. The unweighted coefficient of
variation is calculated using the following mathematical formula:

𝑛
√∑𝑗=1(𝑥𝑖𝑗 −𝑥̅ )/𝑁
CV=
𝑥̅

8|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

where N is the number of districts, in 𝑥𝑖𝑗 , (i) is the socio-economic indicator in the 𝑗 𝑡ℎ region
and 𝑥̅ is mean value of 𝑥𝑖𝑗 .

Table 4 Range of CV (Co-efficient of Variation)

CV Range Disparity Level Interpretation

0-0.05 Low Disparity Regional differences are small; development is

mostly balanced.
0.05-0.2 Moderate Disparity Some inequality exists, but it is not extreme.

0.2-0.6 High Disparity Large differences between regions; some areas are
much more developed.
>0.6 Very High Disparity Extreme inequality; development is highly
unbalanced.

9|Page
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

2.7 Methodological Flowchart

Conceptualization &
Selection of Study Area

Selection of Study
Area

Data Collection

NTL Data (VIIRS-DNB

monthly composite dataset) Socioeconomic Data
(BBS Census, 2022)

ArcGIS (NTL data) Data Processing MS EXCEL (Socio-

economic Data)

Data Analysis

Correlation Analysis Findings Sopher Index, Co-

efficient of Variation

Interpretation of
Results

Final Report

Figure 3 Flow chart of the Methodology (Source: Author, 2025) 10 | P a g e

 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

3. Results
3.1 Correlation Analysis of NTL Data with Socio-economic Indicators.
Table 5 Correlation of All Variables
% Lit Mean Summation of Max Average
Variable Area Intensity Intensity Intensity Correlation
Total Population .591 ** .735** .896** .515** 0.68
Rural Population .336 ** .372** .524** .401** 0.41
**
Urban Population .624 .804** .936** .473** 0.71
Total slum & floating
Population .545 ** .778** .898** .489** 0.68
Financial Account 0.23 .336** .332** 0.067 0.24
Dependency Ratio (Total) 0.382 ** -.373** -.401** -0.04 -0.30
Population growth Rate .692 ** .458** .589** 0.244 0.50
**
Literacy Rate .345 .512** .416** 0.218 0.37
Illetaracy Rate -.345** -.512** -.416** -0.218 -0.37
Main Source of Electricity 0.23 0.202 0.175 0.1361 0.19
Internet User .610** .507** .616** .313* 0.51
**
Pucca Housing .637 .804** .912** .450** 0.70
Semi Pucca Housing .676** .685** .871** .419** 0.66
Kancha Housing 0.15 0.195 .327** 0.237 0.23
Jhupri Housing 0.07 .272* .265* .276* 0.22
**
Foreing Remmitance .369 .335** .529** .390** 0.41
Population Agriculture -0.01 0.047 0.167 0.16 0.09
**
Population industry .791 .796** .944** .440** 0.74
Population service .611** .797** .913** .476** 0.70
Mobile Banking User 0.05 0.191 0.158 -0.17 0.06
**
Own dwelling Unit -.853 -.812** -.892** -.440** -0.75
Not Having Own Dwelling
Unit .853** .812** .892** .440** 0.75
**
Mobile User .380 .460** .411** 0.11 0.34
**Correlation is significant at the 0.01 level (2-tailed).

*Correlation is significant at the 0.05 level (2-tailed).

Table 5 offers a comprehensive breakdown of key societal and economic traits, delving into
their mutual dependencies and influences. It captures details on population spread (across total,
rural, and urban areas), living conditions (divided into sturdy, mixed, or makeshift homes),
financial service availability, and essential amenities like power supply, online access, and
digital banking. The analysis uses indicators such as typical strength, highest strength, total
strength, and lit area coverage to reflect both the norm and the extremes of these features. Some
elements show a clear upward link (like city dwellers and durable homes), while others trend
downward (such as the reliance rate with certain aspects). A central highlight is the mean
linkage score, which reveals the broader connections each trait has with the rest, providing a

11 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

window into their collective impact. Marked symbols flag results with strong statistical weight,
suggesting these patterns are solid across the figures. This set of findings is crucial for decoding
the ties between community growth, housing quality, economic access, and modern facilities
in the region examined.

12 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

3.2 Regional Disparity

3.2.1 Regional Disparity of Literacy-Illiteracy Rate in Bangladesh by Sopher
Index.

Figure 4 Map of Bangladesh: Indicating Regional Disparity of Literacy-

Illiteracy using Sopher Index

13 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Table 6 Range of Regional Disparity by Sopher Index

SI Value Name of Districts Total Characteristics

Number of
Districts
0.41-0.5 Bandarban, Sherpur, Jamalpur 3 Moderate
Disparity
0.5-0.6 Sunamganj, Netrokona, Kurigram 3 Moderate
Disparity
0.6-0.7 Bhola, Khustia ,Meherpur, Gaibandha, Kishoreganj 5 High Disparity

0.7-0.8 Pabna, Rangamati, Mymensingh, Tangail, Sirajganj, 13 Very High

Natore, Rajbari, Manikganj, Chuadanga, Habiganj Disparity
Nilphamari, Lalmonirhat,Rangpur
0.8-0.9 Cox’s Bazar,Khagrachari,Bogura,Magura, 12 Extremely
Shariatpur,Naogaon,Nawabganj,Joypurhat, Disparity
Jehnaidah,Brahmanbaria,Faridpur,Panchagar
0.9-1 Thakurgaon,Rajshahi,Maulivibazar,Lakshmipur, 9 Exceptionally
Madaripur,Noakhali,Satkhira,Narshingdi,Narail High Disparity
1-1.54 Jessore,Khulna,Bagerhat,Jhalokathi,Barisal,Goplaganj, 19 Exceptionally
Khulna,Chittagong,Feni,Chadpur,Munsiganj,Narayanganj, High Disparity
Dhaka,Gazipur,Sylhet,Barguna,Comilla,Pirojpur,Dinajpur
Total=64

The Sopher Index (SI) is a measure of regional inequality, where higher values indicate greater
disparity among regions. A SI value between 0.41 and 0.5 represents moderate disparity, meaning
there are some differences in development, but they are not highly significant. As the SI increases
to 0.5 – 0.6, the disparity remains moderate, though regional imbalances become more noticeable.
When the DI reaches 0.6 – 0.7, the disparity is classified as high, indicating that some regions are
significantly ahead in terms of development compared to others.

A SI range of 0.7 – 0.8 suggests very high disparity, where inequality between regions is severe,
and development is highly uneven. This gap becomes even more extreme when the SI falls
between 0.8 – 0.9, signifying extremely high disparity, where resource distribution and
development opportunities are highly imbalanced. When the SI reaches 0.9 – 1.0, it indicates
exceptionally high disparity, meaning that only a few regions dominate in terms of progress while
others lag far behind. Any SI value exceeding 1.0 further emphasizes exceptionally high disparity,
where inequalities are at their peak, and regional development is heavily skewed.

In summary, a lower SI suggests a more balanced regional distribution, whereas a higher SI

reflects severe inequality and uneven development. Understanding these disparities is crucial for
policy-making, resource allocation, and planning to promote regional equity.

14 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

3.2.2 Regional Disparity of Urban Population in Bangladesh by Co-Efficient

of Variation

Figure 5 Map of Bangladesh: Indicating Regional Disparity of

Population Living in Urban Area using Co-efficient of Variation

The urban population distribution in Bangladesh varies significantly across different regions.
This disparity can be measured using the coefficient of variation (CV), which helps in
understanding the extent of inequality in urban development. Based on CV values, the disparity
levels can be categorized into four groups: low, moderate, high, and very high.

When the CV value is between 0 and 0.05, it indicates low disparity. In this case, regional
differences in urbanization are minimal, meaning most areas have relatively balanced

15 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

development. There are no significant gaps between different regions, and urban growth is
evenly distributed.

A moderate disparity is observed when the CV value ranges between 0.05 and 0.2. Some
inequality exists, but it is not extreme. Certain regions may have slightly higher urbanization
rates than others, but the overall development pattern remains somewhat even.

As the CV increases to 0.2–0.6, the disparity level becomes high. At this stage, significant
differences emerge between regions. Some areas experience rapid urban growth, while others
remain underdeveloped. This leads to a noticeable gap in infrastructure, services, and overall
urban development.

When the CV value exceeds 0.6, it signifies very high disparity. In such cases, extreme
inequality is present, with a few regions dominating urban development while many others lag
far behind. This creates severe imbalances in opportunities, resources, and living conditions
across the country.

Understanding these disparity levels is crucial for policymakers. By analyzing CV values, they
can identify underdeveloped areas and implement targeted strategies to promote balanced
urbanization in Bangladesh.

16 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

4. Discussion and Conclusion

The Table 5 shows the correlation between various socio-economic factors and four measures
of night-time light intensity: % Lit Area, Mean Intensity, Summation of Intensity, and Max
Intensity. The values range from -0.85 to 0.94, indicating how different factors influence
illumination levels across regions.

Population-related variables exhibit strong correlations with night-time light. Total Population
shows a high correlation, ranging from 0.515 (Max Intensity) to 0.896 (Summation of
Intensity), confirming that densely populated areas tend to be brighter. Urban Population has
an even stronger relationship, with the highest correlation reaching 0.936 (Summation of
Intensity), while Rural Population shows much weaker correlations, with values between 0.336
and 0.524. Similarly, Slum & Floating Population maintains strong positive correlations,
reaching 0.898, suggesting that even informal settlements contribute significantly to night-time
brightness.

Economic and infrastructural factors also play a crucial role. The Industrial Population has one
of the highest correlations, ranging from 0.440 to 0.944, highlighting the concentration of
artificial light in industrial zones. The Service Sector Population follows a similar trend, with
values between 0.476 and 0.913, emphasizing the impact of commercial activities on
illumination. Foreign Remittances show moderate correlations, peaking at 0.529, indicating
that regions receiving more financial inflows tend to be more developed and well-lit.

Education and technological factors demonstrate notable correlations with night-time light.
Literacy Rate has a positive relationship, with correlations between 0.218 and 0.512, while
Illiteracy Rate exhibits a negative correlation, ranging from -0.218 to -0.512, reinforcing the
link between education and development. Internet Users show a moderate correlation with light
intensity, reaching up to 0.616, while Mobile Users have a weaker correlation, peaking at
0.460. This suggests that digital connectivity is associated with more urbanized and developed
areas that experience higher illumination.

Housing characteristics reveal interesting trends. Pucca Housing (permanent structures) has
strong positive correlations, ranging from 0.450 to 0.912, confirming that well-built housing is
linked to higher night-time light intensity. Semi-Pucca Housing follows a similar trend, with
correlations between 0.419 and 0.871. In contrast, Kancha Housing and Jhupri Housing
(temporary or makeshift structures) show much weaker correlations, suggesting that these
housing types are found in less illuminated areas. A particularly striking finding is the strong
negative correlation of Own Dwelling Unit, which ranges from -0.440 to -0.892. This suggests
that areas with high homeownership rates, likely rural or suburban regions, tend to have lower
night-time illumination. Conversely, Not Having an Own Dwelling Unit has an equally strong
positive correlation (0.440 to 0.892), implying that rental or temporary housing is more
common in brightly lit urban areas.

Other socio-economic factors show varying correlations with night-time illumination. The
Dependency Ratio has a mixed relationship, with a weak positive correlation of 0.382 for %
Lit Area but negative correlations for the other three measures, as low as -0.401. This indicates
that regions with higher dependent populations tend to have lower night-time light intensity.
Financial Account Ownership shows weak correlations, between 0.067 and 0.336, suggesting
that banking access alone does not strongly influence illumination levels. Agricultural

17 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Population has very weak correlations, ranging from -0.01 to 0.167, reinforcing that farming
regions are less illuminated compared to industrial and service-oriented areas.

In summary, the findings highlight that urbanization, industrialization, education, and

infrastructure development are the strongest drivers of night-time light intensity, with
correlations reaching as high as 0.94. In contrast, rural settings, high dependency ratios, and
high homeownership rates are linked to lower illumination, with negative correlations reaching
-0.85. This reinforces the idea that economic activity, urban living, and technological
advancement are closely tied to brighter night-time environments.

The Table 6 represents regional disparity in literacy and illiteracy across Bangladesh using the
Sopher Index (SI) instead of the Disparity Index (DI). The Sopher Index is used to measure the
extent of literacy inequality among districts, highlighting variations in educational
development.

Districts with moderate disparity (SI: 0.41–0.6), such as Bandarban, Sherpur, Jamalpur,
Sunamganj, Netrokona, and Kurigram, show some educational inequality, but the gap between
literate and illiterate populations is not extreme. High disparity (SI: 0.6–0.7) is found in Bhola,
Khustia, Meherpur, Gaibandha, and Kishoreganj, where literacy rates vary significantly, and
access to education is unevenly distributed.

Very high disparity (SI: 0.7–0.8) is seen in Pabna, Rangamati, Mymensingh, Tangail, Sirajganj,
and others, where some areas have better literacy rates, while others remain educationally
backward. Extremely high disparity (SI: 0.8–0.9), observed in Cox’s Bazar, Khagrachari,
Bogura, and Magura, indicates a severe literacy gap, with some districts struggling with high
illiteracy rates.

The highest disparity (SI: 0.9–1.54) is found in Thakurgaon, Rajshahi, Khulna, Chittagong,
Dhaka, and others, where urban centers have significantly higher literacy rates than rural areas,
leading to a serious educational imbalance. To reduce these disparities, policy interventions,
improved access to education, and literacy programs are essential for ensuring equal
educational opportunities across all regions.

The Coefficient of Variation (CV) was used as a statistical measure to examine regional
disparities in development across districts. This analysis provides insights into how evenly or
unevenly development indicators are distributed geographically. Based on the CV values, the
districts were classified into four categories: low, moderate, high, and very high disparity
shows in Table 4.

A total of 22 districts falls under the low disparity category (CV between 0 and 0.05), indicating
that development in these areas is relatively uniform. The small variation in CV values suggests
that these districts enjoy a balanced distribution of development outcomes, with no significant
differences in access to infrastructure, services, or economic opportunities.

The majority of the districts — 39 in total — are categorized as having moderate disparity (CV
between 0.05 and 0.2). This points to the presence of some inequality in these regions, although
it is not extreme. While development is not perfectly balanced, the disparities are within a range
that is still considered manageable. These districts might benefit from targeted interventions to
prevent further widening of regional gaps.

18 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Only one district was identified as experiencing high disparity (CV between 0.2 and 0.6). This
indicates a substantial imbalance in development, where parts of the district may be
significantly more advanced than others. Such disparity can lead to unequal access to resources
and opportunities, which may eventually hinder overall progress if not addressed.

Finally, 2 districts were found to have very high disparity, with CV values exceeding 0.6. These
areas are characterized by extreme inequality, where development is highly concentrated in
certain pockets while other areas remain significantly underdeveloped. Such stark contrasts can
have serious social and economic implications and highlight the need for immediate, focused
policy interventions to reduce the imbalance.

The multifaceted analysis undertaken—encompassing night-time light intensity, educational

disparity via the Sopher Index, and regional development variation through the Coefficient of
Variation—reveals a pronounced spatial asymmetry in Bangladesh’s socio-economic
landscape. The illumination metrics, strongly aligned with urbanization, industrialization, and
infrastructural sophistication, underscore the pivotal role of economic agglomeration in
shaping spatial development patterns. The robust positive correlations between night-time light
and variables such as urban population, industrial labor force, internet penetration, and durable
housing typologies, highlight the luminance of urban cores as epicenters of growth and
opportunity.

Conversely, the negative associations observed with rurality, agricultural dependence, and high
homeownership rates—often indicative of subsistence-based economies—reflect the persistent
developmental inertia characterizing peripheral regions. The disparity in educational
attainment, quantified through the Sopher Index, further accentuates this spatial disequilibrium.
Districts exhibiting high to extreme levels of educational disparity reflect not merely unequal
access to literacy but also the enduring structural and systemic impediments to equitable human
capital formation.

The CV-based classification of regional disparity further validates these observations,

demarcating a developmental gradient wherein a select few districts exhibit highly polarized
growth trajectories. Such concentration of resources and opportunities in limited geographic
locales exacerbates inter-regional inequality and perpetuates a cycle of uneven development.

Collectively, these empirical insights illuminate the urgent imperative for spatially inclusive
policy architectures. Strategic interventions—anchored in regional planning, educational
equity, and rural economic revitalization—are essential to attenuate existing disparities.
Without deliberate, targeted action, the entrenchment of socio-economic dualism between the
urban-industrial enclaves and the rural hinterlands will only intensify, thereby impeding the
nation’s aspiration for holistic and sustainable development.

19 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Credit authorship contribution statement

211702- Data cleaning, formatting, Software (MS-Excel)

211704-Writing-review & editing, Writing – original draft, Data analysis, Methodology,

Software (ArcGIS, MS-Excel, SPSS etc.), Statistical computations.

211716- Helped organize the Sopher Index data, Software (MS-Excel).

211732- Worked on data tabulation and preliminary statistical summaries.

211737- Cross-checking data consistency across sections, Software (MS-Excel).

211741- Processing of satellite imagery using Google Earth Engine, spatial analysis using
ArcGIS, including mapping and calculation of disparity indices, Software (MS-Excel).

Special Thanks- Tanmoy Mazumder - Supervision, Conceptualization.

References
Jean, N., Burke, M., Xie, M., Davis, W. M., Lobell, D. B., & Ermon, S. (2016). Combining satellite
imagery and machine learning to predict poverty. Science, 353(6301), 790–794.
https://doi.org/10.1126/science.aaf7894

Lazar, M. (2011). Shedding Light on the Global Distribution of Economic Activity. The Open
Geography Journal, 3(1), 147–160. https://doi.org/10.2174/1874923201003010147

Rahman, M. S., Mohiuddin, H., Kafy, A. Al, Sheel, P. K., & Di, L. (2019). Classification of cities in
Bangladesh based on remote sensing derived spatial characteristics. Journal of Urban
Management, 8(2), 206–224. https://doi.org/10.1016/j.jum.2018.12.001

Sangkasem, K., & Puttanapong, N. (2022). Analysis of spatial inequality using DMSP-OLS nighttime-
light satellite imageries: A case study of Thailand. Regional Science Policy and Practice, 14(4),
828–849. https://doi.org/10.1111/rsp3.12386

Wahed, M., Rizvee, R. A., Haque, R. R., Ali, A. M., Zaber, M., & Ali, A. A. (2020). What Can Nighttime
Lights Tell Us about Bangladesh? 2020 IEEE Region 10 Symposium, TENSYMP 2020, June, 1612–
1615. https://doi.org/10.1109/TENSYMP50017.2020.9230806

Wasim, S., & Munir, K. (2017). Regional Disparity and Decentralization in Pakistan: A Decomposition
Analysis. MPRA Paper, No.83444.

20 | P a g e
 Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

Zeng, Q., & Zhang, W. (2019). Research on the Development of “Ghost City” Based on Night Light
Data: Taking Sichuan Province as an Example. Open Journal of Social Sciences, 07(12), 176–188.
https://doi.org/10.4236/jss.2019.712013

BBS. (2022). Population and Housing Census. Bangladesh Bureau of Statistics.

21 | P a g e
Nighttime Light Derived Assessment of Regional Inequality of Socioeconomic Development in Bangladesh

22 | P a g e