Available via license: CC BY 4.0
Content may be subject to copyright.
DATA DRIVEN FRAMEWORK FOR ANALYSIS OF AIR QUALITY LANDSCAPE FOR
THE CITY OF LAHORE
A. Rahman*, M. Usama, M. Tahir, M. Uppal
Lahore University of Management Sciences (LUMS), Pakistan
(abdur.rahman, muhammadusama, tahir, momin.uppal)@lums.edu.pk
Commission IV, WG IV/9
KEY WORDS: Air Quality, Air Quality Monitoring System, Low Cost Air Quality Monitoring Stations, Air Quality Data
ABSTRACT:
Several Pakistani cities are among the world’s most polluted. In the previous three years, air pollution in Lahore has been con-
siderably over World Health Organization guideline levels, endangering the lives of the city’s more than 11 million citizens. In
this paper, we investigate the city’s capability to combat air pollution by analyzing three essential aspects: (1) Data, (2) Capacity,
and (3) Public awareness. Several studies have reported the need for expansion of the current air quality monitoring network. In
this work, we also provide a context-aware location recommendation algorithm for installing the new air quality stations in Lahore.
Data from four publicly available reference-grade continuous air quality monitoring stations and nine low-cost air quality measuring
equipment are also analyzed. Our findings show that in order to measure and mitigate the effects of air pollution in Lahore, there
is an urgent need for capacity improvement (installation of reference-grade and low-cost air quality sensors) and public availability
of reliable air quality data. We further assessed public awareness by conducting a survey. The questionnaire results showed huge
gaps in public awareness about the harms of the air quality conditions. Lastly, we provided a few recommendations for designing
data-driven policies for dealing with the current apocalyptic air quality situation in Lahore.
1. INTRODUCTION
Air pollution is defined as the contamination of indoor or out-
door air with chemical, physical or biological agents modifying
the characteristics of air. Most common sources of outdoor air
pollution include exhaust combustion from motor vehicles, in-
dustrial emissions, forest fires, livestock farming, fertilizer, and
power plants. Major constituents of air pollution are Particu-
late Matter (PM), Carbon Monoxide (CO), Ozone (O3), Nitrous
(NOx), and Sulfur oxide (SOx). PM includes fine particles
suspended in the air. These particles are usually 2.5 micro-
meter (PM2.5) and 10 micrometer (PM10) in diameter (Vallero,
2014). These particles are a byproduct of combustion in mo-
tor vehicles, burning fossil fuels, industrial processes, and other
sources of smoke. Other indirect sources are different chem-
ical reactions of NOxand SOxin air. Major health effects of
the PM pollutants are decreased lung function, eye, nose, and
throat irritation, difficulty breathing, aggravated asthma, non-
fatal heart attacks, and even premature deaths in people with a
lung or heart disease history.
Particulate matter is a significant contributor to air pollution.
Air quality is measured using air quality monitoring systems
(AQMS), which are technically validated by organizations such
as the United States Environmental Protection Agency (EPA),
among others, and are also known as reference-grade AQMS.
Reference-grade AQMS are costly and require a significant
amount of upkeep. The PM2.5values recorded by these
reference-grade stations are regarded authentic and are utilized
by appropriate authorities to issue health advisories. In most
developed and underdeveloped countries, there is a severe gap
in the installed reference-grade AQMS resulting in huge cover-
age gaps and inconsistent air quality data. Considering sparse
∗Corresponding author
resources and the overburdened economies of developing coun-
tries, expansion of the current air quality measurement network
in a short time is a challenge. On the other hand, low-cost al-
ternatives exist, but the trustworthiness of their reported values
is frequently questioned. Many studies have been conducted to
enhance the performance of low-cost sensors. According to the
literature, a mix of reference-grade AQMS and low-cost sensors
can aid in the development of urban city-scale measurement
networks while keeping the economic aspects of developing na-
tions in check (Gulia et al., 2015, Usama et al., 2022).
Air pollution has emerged as a significant issue in the subcon-
tinent. Pakistan has recently seen a yearly ”smog” season that
lasts from November to February each year. Multiple Pakistani
cities have made the list of the world’s most polluted cities in
recent years. Lahore, the provincial capital of Punjab, is one
of the world’s three most polluted cities. The current state of
Lahore’s air quality puts the lives of the city’s 11 million resid-
ents1in grave danger. For the most part of the last three years,
the air quality index (AQI) stayed between poor to severe. The
AQI is a metric used to quantify the effect of air pollution on hu-
man health based on limited exposure. The higher the AQI, the
more health risks there are. In winters the smog, fog, and haze
results in the closure of the major highways, airports, and trans-
portation incurring economic losses and social unrest. Many
road accidents due to smog are also reported resulting in deaths
and financial losses.
Conditions deteriorated to the point where the government was
compelled to take action to protect the public from additional
exposure and pollution. It includes the closure of brick kilns
in 2018 and the enforcement of conventional brick kiln conver-
sion to zig-zag technology with lower air pollutants (Mukhtar,
1https://www.pbs.gov.pk/content/
final-results- census-2017
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
167
2018). In recent years, the government has also issued orders
to shut down industries during the pollution season, which has
had a disastrous effect on the economy. In 2021, Punjab govern-
ment also decided to close every Monday for nearly a month to
combat smog2. These circumstances need a thorough examina-
tion of the city’s air quality landscape. In this study, we focused
on the available PM2.5data sources in Lahore, Pakistan, and in-
vestigated the capability of the city to handle the issue of poor
air quality.
Air quality and a city’s capability to tackle air pollution are
quantified using three indicators: (1) capacity, (2) data, and (3)
public awareness. Here, capacity refers to public and private
measurement infrastructure, data refers to the public availabil-
ity of air quality measurements to develop data-driven policies,
and public awareness refers to the general public’s interest in
the issue and how the public views the ramifications of air qual-
ity issues. In this paper, we make the following analysis and
contributions
1. We have collected and prepared a dataset of the
PM2.5 measurement data from various publicaly available
sources (reference-grade AQMS and low-cost sensors) and
analyze it to further reflect on the robustness and authenti-
city of the reported data from the public sources.
2. Based on the prepared dataset and the context information
of Lahore city, we have developed an algorithm for recom-
mending deployment positions of new air quality measure-
ment sensors. We have also reflected upon the validity of
these results and how more context information can yield
better sensor deployment.
3. Perception versus reality plays a vital role in swaying the
opinion of the urban public to adopt better practices for en-
suring prevention against the hazardous effects of air pol-
lution. We have conducted a survey to gauge perception
vs. reality of the air quality in Lahore, and this work also
provides the crux of these findings.
The remainder of the paper is structured as follows. In the next
section, we will provide specifics on the air pollution data for
Lahore city as well as the sources. Section 3 describes the pro-
posed context-based location recommendation approach for air
quality sensor placement and the results achieved. Section 4
will discuss the survey conducted during this location recom-
mendation approach to determine how individuals view the is-
sue of air quality in Lahore. Section 5 contains a brief yet
informative discussion on the validity of the acquired results.
Finally, in Section 6, the work is concluded, along with a dis-
cussion on potential future directions.
2. AIR QUALITY DATA OF LAHORE
Needless to say, the poor air quality in the city of Lahore has
significantly adverse health effects on its residents. Life Index
(AQLI) from the Energy Policy Institute at the University of
Chicago (EPIC) released its 2021 annual report on the effects
of poor air quality on the health of an average Pakistani citizen
(Greenstone and Fan, 2018, Greenstone and Fan, 2020). The
AQLI report suggests that an average Pakistani tends to lose
3.9 years of life expectancy if the current levels of air pollution
2http://sdsa.lums.edu.pk/GrandChallengeFund/
BlogArchive/3
persist (Greenstone and Fan, 2020). It also reports that in a
few most polluted areas, the loss in life expectancy can go up
to 7 years (Greenstone and Fan, 2020). An average citizen of
Lahore is expected to gain 5.3 years’ worth of life expectancy
if the World Health Organization (WHO) guidelines for PM2.5
are met.
Environmental protection departments/agencies often collect
air quality data, and health advisories are issued based on
pre-defined AQI readings. Pakistan, like any other develop-
ing/underdeveloped country, does not have an adequate number
of AQMS installed, thus there is a scarcity of available air qual-
ity data for most of the country. Only four reference-grade air
quality monitors and a few low-cost sensors with publicly avail-
able air quality data are available in Lahore. We have confined
our investigation to PM2.5because air quality data for majority
of Lahore is not accessible for other criteria pollutants. Figure
1 provides the location of the reference-grade and low-cost air
quality monitors in Lahore city.
Figure 1. There are four reference-grade and eight low-cost air
quality monitoring stations in Lahore. The figure depicts the
location of the sensors on the map of Lahore.
Three reference-grade sensors in Lahore are from Environ-
mental Protection Department (EPD), Punjab3, and one from
US consulate4 air quality station. The data from EPD is avail-
able at daily granularity in PDF format. The PDF files were
digitized using python-based tools as well as manual scrapping.
The data from US Embassy was available at hourly granularity.
We have converted both sources to daily granularity for com-
parison and depicted the data timeline in Figure 2. There are
eight low-cost sensors deployed by “PurpleAir”5 with publicly
available PM2.5 concentration values in Lahore. We have also
converted them in daily granularity, and Figure 3 depicts the
timeline of the data availability. Here we also want to note that
there are a few other low-cost sensors in Lahore deployed by
IQAir6 but their recorded data is not publicly available.
We have noticed that two of the reference-grade stations
stopped reporting PM2.5 values in mid-2020 (Dental college
station) and mid-2021 (Met station) respectively. Only two
reference-grade AQMS are reporting PM2.5 concentrations for
approximately 670 Square kilometers which is not acceptable
3 https://epd.punjab.gov.pk/aqi
4 https://www.airnow.gov/international/
us-embassies- and-consulates/#Pakistan$Lahore$
5 https://www2.purpleair.com/
6 https://www.iqair.com/
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
168
Figure 2. Timeline plot of PM2.5 data for reference grade
sensors
Figure 3. Timeline plot of PM2.5 data for low-cost grade sensors
by any stretch of the imagination. All low-cost sensors came
online in mid-August 2021. Their calibration and ability to
cater to context while reporting data lacks credibility.
3. AIR QUALITY MONITORING NETWORK OF
LAHORE
In a recent verdict7, the Lahore High Court ordered the Punjab
government to take concrete steps to curb the air pollution in
Lahore. The court also stated that corporations should also ful-
fill t heir corporate social responsibility and reduce emissions,
and government must enforce the already existing suggestions
from the Lahore smog commission report and Punjab clean air
act to mitigate the smog situation. The Lahore smog commis-
sion report8 recommended that the Punjab government must in-
crease the active reference-grade air quality monitoring stations
from three to twelve. Based on the smog commission report,
EPD Punjab needs to install nine more reference-grade AQMS.
Here the challenging bit is to determine the optimal location
for the installation of AQMS. Several factors need to be con-
sidered before deciding on the optimal installation location. A
few of these factors are namely; availability of land, possible
conflicts w ith f uture u rban c onstruction a nd e xpansion plans,
accessibility to maintenance staff and other services, security
of the site, and some heuristic constraints developed by previ-
ous experiences, etc. Despite the fact that these conditions are
strong administrative markers for the installation, research on
air quality sensor placement suggests that a context-aware data-
driven method can give optimal installation locations.
Sun et al. proposed a citizen-centric air quality sensor place-
ment technique, where they have used Cambridge city traffic
patterns, point of interest values, and demographic statistics as
context information. They modelled the location recommend-
ation as a linear integer programming model in which both the
objective function (location) and the constraints (context) are
considered to be linear, which we believe will result in a high
7 https://data.lhc.gov.pk/reported_judgments/green_
bench_orders
8 \footnote{\url{https://epd.punjab.gov.pk/system/
files/Smog\%20commission\%20report.pdf}}
rate of false positive locations as the number of sensors (to be
placed) increases (Sun et al., 2019). Kelp et al. presented a
PM2.5sensor placement approach where they used multiresol-
ution dynamic mode decomposition (mrDMD) on 16 years of
historical PM2.5data to suggest new sensor location (Kelp et
al., 2022). Zhou et al. compared five sensor placement tech-
niques (random, minimization of the matrix condition num-
ber for sensor placement, empirical interpolation method for
sensor placement, local extrema-based techniques, and QR-
factorization method for sensor placement) from the control
theory and fluid dynamics literature on a couple of satellite-
derived huge PM2.5datasets from China (Zhou et al., 2022).
Though the compared techniques are well embedded in the lit-
erature, we do not have this sort of high-resolution spatiotem-
poral dataset available for Lahore city. Mohar et al. used QR
factorization, singular value decomposition (SVD), and ma-
chine learning-based techniques to design an optimal sensor
placement technique for signal reconstruction in control sys-
tems (Manohar et al., 2018). Though this technique has shown
promise in sensor and actuator placement given the small num-
ber of sensors (in our case 09) and limited untrustworthy sparse
historical data the technique is highly likely to produce false
positives. There are a few other optimal sensor placement tech-
niques from communication systems and traditional sensors
network literature (Younis and Akkaya, 2008) but these tech-
niques are not optimal for a gappy, untrustworthy, sparse, and
small dataset to determine the best locations for sensor place-
ment in Lahore city.
Microsoft’s Urban Air Project9is at the cutting edge of sensor
placement research for air quality measurements. They have
developed a method that uses attributes from the previously in-
stalled sensors and historical air quality data from existing sta-
tions to suggest suitable locations for future air quality stations
(Hsieh et al., 2015). Hsieh et al. used an affinity graph-based
technique to determine the optimal location for AQMS place-
ment for Bejing city in China. The proposed procedure also
incorporates the historical PM2.5concentrations, meteorology
data, road network, POI data, etc., to ensure that appropriate
context is also incorporated in the optimal location selection
(Hsieh et al., 2015), (Zheng et al., 2013). The technique pro-
posed in this paper is an extension of the affinity graph-based
approach where we have included various context features col-
lected from Lahore city and used whatever historical air quality
data is available.
3.1 Affinity Graph Based Location Recommendation Sys-
tem
Hsieh et al. used affinity graphs with a greedy entropy min-
imization model to develop a location recommendation system
for AQMS installation location recommendation (Hsieh et al.,
2015). Initially, the city is divided into graph nodes and edges.
Where every edge has an associated weight, and every node
has an associated set of features (road networks, residential
areas, commercial areas, industrial areas, public spaces, met-
eorological features, and other factors that may contribute to
variation in air pollution). Since most location recommenda-
tion techniques are designed and tested for developed countries
with historical data on air pollution and related context features
are publicly available, the location recommendation for the new
AQMS becomes a simple task. Since we are trying to predict
the location of the AQMS for an underdeveloped/developing
9https://www.microsoft.com/en-us/research/project/
urban-air/
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
169
Figure 4. Flow chart of the station location recommendation algorithm.
city where the available historical data is sparse and untrust-
worthy and context features are also not available, the simple
learning task becomes a real hassle.
For Lahore, we have collected historical air quality data from
all known publicly available sources. The data was cleaned and
preprocessed by following the data science principles (clean-
ing, normalization, outlier detection, etc). We have combined
the reference-grade AQMS data with low-cost sensors data to
further improve the volume of the dataset. Gathering the con-
text data for Lahore city is challenging as to the best of our
knowledge, it is not publicly available. We have employed Geo-
graphic Information Systems (GIS) tools for collecting context
information such as commercial hubs, industrial areas, traffic
hotspots, etc. The meteorology data was available. The road
network was extracted using satellite imagery and machine
learning techniques. Figure 5 depicts the identified commercial
hubs, industrial areas, drainage streams, and traffic hotspots.
Figure 5. Identified pollution hotspots in Lahore
Since the proposed method is inspired from (Hsieh et al., 2015),
we recommend the interested reader to see (Hsieh et al., 2015)
for in-depth details on how affinity g raphs c an b e leveraged
for designing location recommendation algorithms. Our model
works in two stages. The first step is to compute the probability
distribution of unlabeled nodes using feature weight matrices,
graph weight matrices, and labeled nodes. The entropy for each
node is computed in the second phase, and the node with the
lowest entropy is marked as the labeled node and given the low-
est rank for a recommendation for the installation of a new air
quality station. The model is given a new set of labeled and
unlabeled nodes iteratively. A more detailed description of the
AQMS location inference method in the following steps:
1. The input to the proposed location inference technique is
labeled node list, labeled node air quality values (histor-
ical), context features associated with each node, unla-
belled node list (candidate node locations), and features
associated with the unlabelled candidate nodes.
2. Based on the input, an affinity graph is created, and graph
weight initialization is performed (in our case, we have
initialized it with all 1’s).
3. After the initialization, we have computed the initial en-
tropy with the uniform underlying distribution of un-
labeled nodes, and the feature weight matrix is updated
using gradient descent. The probability distribution of the
updated unlabelled nodes is estimated using a harmonic
function, and the entropy of the updated graph is com-
puted.
4. The difference between the entropy of pre and post-
graph updates is computed and subjected to a difference
threshold. If the entropy difference is less than the pre-
defined threshold value, we ranked the node with the low-
est entropy in the reverse order and assigned the node to a
labeled node (predicted AQI assignment). Whereas, if the
entropy difference is greater than the predefined threshold,
the algorithm will go back and update the graph, and fea-
ture weights and entropy is recomputed.
5. If there is no unlabeled node left in the graph, the proposed
algorithm will output the “N” highest-ranked nodes, which
in our case are the first nine nodes. Since we have conver-
ted Lahore into a graph grid, the label of the top “N” nodes
will also provide their location information on the map.
6. We observed an inherent problem of clustering multiple
recommended nodes in this algorithm. We noticed that
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
170
the conversion of unlabeled nodes from near the labeled
nodes converged to the point where none of the nodes
were labeled because the node with the lowest entropy was
tagged as a labeled node. It is an issue when the number of
recommended locations is more than one. Thus multiple
nodes were recommended in a cluster. To tackle this prob-
lem, we introduced another loop to recommend only one
location in every iteration and use the previous recommen-
ded location as a labeled node in the next iteration.
The flow chart for the algorithm used for inferring the optimal
AQMS placement is depicted in Figure 4.
We tested the proposed algorithm with the available air quality
data for Lahore city and the collected context features. The pro-
posed AQMS location recommended algorithm provided 09 re-
commended for new AQMS deployment. The output of the pro-
posed model for the recommendation of 09 stations for Lahore
city is provided in Figure 6. Reference-grade AQMS are usually
deployed by the environmental protection department/agencies,
which EPD Punjab will do in the future, though for now, we are
planning to deploy 09 low-cost on the recommended locations
to collect PM2.5concentrations. In future work, we intend to
analyze the efficiency of the proposed location algorithm using
the reported data from all air quality measurement sources.
Figure 6. Recommended locations for installation of air quality
stations (yellow) with the locations of currently installed stations
(red)
4. PUBLIC AWARENESS
The third component of accessing the ability of a city to deal
with the air quality issue is its public awareness. Following the
tradition in the literature (Liu et al., 2017, Pantavou et al., 2017,
Lou et al., 2022, Maione et al., 2021), we have designed a sur-
vey with only ten simple questions addressing the perception of
the air quality among citizens and how this perception is built,
and whether they are taking any measures to avoid detrimental
consequences of the air pollution in Lahore city. The question-
naire varied from general public perception of air pollution and
public information sources to the self-reported levels of health
impact due to air pollution.
We have conducted this survey from 21st December 2021 to
10th January 2022. We have received 177 responses from dif-
ferent universities in Lahore. As expected, nearly 60% of the
participants were between 18-22 years old. 67% of the re-
sponses came from males, and 37% from female participants.
We collected responses from 11 different public and private uni-
versities in Lahore. Nearly 73% of the responses suggested that
the air quality in Lahore ranges from very poor to severe (44%
said severe and 29% said very poor). Nearly 81% of the re-
sponses suggested that smog/fog/haze is their sensory percep-
tion for air quality. This response statistics also show that many
people think that air pollution only exists in winter, and as soon
as the smog is gone, the air quality is back to normal, which is
not the case. This response also suggests the lack of information
about air pollution among young people. Our result also shows
that news, mobile applications, and social circles are domin-
ant ways to get air quality-related information. Nearly 15% of
people responded that they do not take any precautions to deal
with air pollution. 31% of the total responses suggested that
they have a respiratory condition, and out of those 31% people,
40% said that their condition aggravated due to poor air quality
in Lahore. Lastly, 38% of people suggested that transport is the
major contributor to poor air quality in Lahore. The rest of the
people responded in favor of industries, agriculture, etc. The
responses distribution about the potential sources of the survey
are depicted in Figure 7. More details on our survey and results
are available on our blog-post10.
This survey suggests the need to raise public awareness about
the dangers of air pollution, and we suggest that including com-
ponents on air pollution, its causes, and ways to deal with it
must be included in the school, college, and university cur-
riculum. We also recommend public meetings, town halls, and
seminars to increase awareness among people. We also sug-
gest the give incentives to the residential and commercial areas
where the air quality improves.
Figure 7. Public awareness survey response on the perception
about the potential air pollution sources in Lahore.
5. DISCUSSION AND RECOMMENDATIONS
In this article, we assessed Lahore’s ability to cope with the
growing air pollution problem in three areas: data, capability,
and public awareness. Our findings show that Lahore’s present
air quality monitoring network has significant gaps, and the city
is unable to deal with the ever-increasing menace of air pollu-
tion. Lahore does not have enough AQMS installed, and the
available AQMS data is gappy, unreliable, and does not reflect
the severity of the air pollution. We suggest that EPD Punjab
needs to increase the reference-grade AQMS and also support
10 http://sdsa.lums.edu.pk/GrandChallengeFund/
BlogArchive/4
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
171
the deployment of low-cost sensors. EPD also has to ensure
an ample amount of calibration and chemical analysis facilities
to further the quality of sensing and source apportionment. The
development of a public dashboard is prevalent where all public
and private data on the air quality is gathered and analyzed and
made available to the public. There is a dire need for academic
and industrial partners to further enhance the agenda on gath-
ering reliable air quality data and using machine learning and
other predictions technique to prepare the EPD for upcoming
air quality challenges.
In this work, we also did our bit by designing an AQMS place-
ment algorithm that incorporates the local context to provide
optimal location. This effort is also a candidate solution to effi-
ciently using the limited financial resources available. Here we
also want to note that deploying 09 AQMS in Lahore will not
solve the data availability issue, general public, housing societ-
ies, and institutions also have to play an active role in the dens-
ification of the air quality monitoring network. We recommend
that all government and private housing societies, universities,
industries, hospitals, etc., install AQMS (near reference or low-
cost) and make the date available to the public and government.
Public awareness is the third component of accessing the abil-
ity of Lahore city to cope with the air quality issues. Our sur-
vey indicates the lack of awareness about the air quality issues.
Our survey also reveals disparities between public perception
and real air quality. There is an urgent need to raise public
knowledge about air quality through awareness campaigns and
community activities. The general public is unaware of the vari-
ous resources for reporting air quality. More emphasis should
be placed on preventative measures such as wearing masks, in-
stalling air purifiers, and reducing outside activity during pollu-
tion. Studies on source apportionment may aid in quantifying
the sources of air pollution in Pakistan. We recommend includ-
ing air quality and associated information in school, college,
and university curricula, collaborating with religious experts to
emphasize the issue of air quality in Friday sermons and other
religious gatherings, supporting public awareness events such
as seminars and town halls, and offering incentives to the resid-
ential and commercial areas where the air quality improves.
5.1 Challenges
There are a few significant obstacles and trade-offs that devel-
oping/underdeveloped nations must confront, which necessit-
ate a concerted global effort to address the air pollution prob-
lem. Following are a few challenges (keeping in view the un-
derdeveloped/developing countries):
1. Data collection and public datasets: Collecting air qual-
ity data is a difficult process since varying concentrations
of air contaminants are involved. Given the environmental
and health dangers associated with poor urban air qual-
ity, it is critical to creating a centralized real-time air qual-
ity data monitoring and processing system. Sensor-centric
data collection and crowd-centric data collection are two
approaches for obtaining urban data (e.g., air quality data,
POI, meteorological data, etc.). There are two types of
sensor-centric paradigms. These classifications are based
on whether the sensors gathering data are mobile (de-
ployed on a moving item) or static (deployed on a fixed
location). There are two types of crowd-sourced data col-
lection: active (data created through participation surveys
and check-ins) and passive (data generated by users pass-
ively while using the urban infrastructure). It is important
to understand what type of approach can help gather the
most reliable and larage amount of air quality data (Usama
et al., 2022).
2. Trade-off between economic growth and air pollution:
Most developing countries are trying to manage their eco-
nomies, and the balance between economic expansion and
air pollution is almost always skewed toward economic
growth. Finding a middle ground between economic
growth and air pollution is a challenging task. Health and
environmental budgets are diminishing, making it difficult
for developing countries to detect, report, and improve the
air quality (Xu and Li, 2018, Usama et al., 2022).
3. Regularization and air quality measurements: Un-
derdeveloped/developing countries must implement data-
driven policies, with regularisation based on data and local
context. Once these policies are developed, the adminis-
tration must guarantee that they are executed (Yamineva
and Romppanen, 2017, Usama et al., 2022).
4. Public awareness: Unfortunately, in under-
developed/developing nations, public understanding
of the hazards of air pollution and efforts to ameliorate its
consequences is relatively low. With the development of
social media applications and the Internet penetration, the
government may easily overcome this challenge. The ad-
ministration should use social media platforms, television
shows, print and digital platforms, town halls, seminars,
hackathons, conferences, and so on to disseminate as
much information as possible about the effects of air
pollution on human health and the local economy, major
air pollution sources, and what options the public has to
help reduce emissions.
6. CONCLUSIONS
In this article, we assessed Lahore’s potential to deal with the
looming challenges of air pollution across three dimensions:
data, capacity, and public awareness. The goal of gaining ac-
cess to Lahore’s ability to deal with air quality on the afore-
mentioned verticals is to establish the groundwork for building
a framework for developing an air quality network for under-
developed countries. We also proposed an optimal placement
suggestion approach for installing air quality monitoring sta-
tions. Finally, we have identified a few challenges that develop-
ing nations must solve in order to avert the air apocalypse.
REFERENCES
Greenstone, M., Fan, C. Q., 2018. Introducing the air quality
life index. AQLI Annual Report.
Greenstone, M., Fan, Q. C., 2020. Pakistan Fact Sheet.
Gulia, S., Nagendra, S. S., Khare, M., Khanna, I., 2015. Urban
air quality management-A review. Atmospheric Pollution Re-
search, 6(2), 286–304.
Hsieh, H.-P., Lin, S.-D., Zheng, Y., 2015. Inferring air quality
for station location recommendation based on urban big data.
Proceedings of the 21th ACM SIGKDD international confer-
ence on knowledge discovery and data mining, 437–446.
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
172
Kelp, M. M., Lin, S., Kutz, J. N., Mickley, L. J., 2022. A new
approach for determining optimal placement of PM2. 5 air qual-
ity sensors: case study for the contiguous United States. Envir-
onmental Research Letters, 17(3), 034034.
Liu, H., Kobernus, M., Liu, H. et al., 2017. Public perception
survey study on air quality issues in Wuhan, China. Journal of
Environmental Protection, 8(10), 1194.
Lou, B., Barbieri, D. M., Passavanti, M., Hui, C., Gupta,
A., Hoff, I., Lessa, D. A., Sikka, G., Chang, K., Fang, K.
et al., 2022. Air pollution perception in ten countries during
the COVID-19 pandemic. Ambio, 51(3), 531–545.
Maione, M., Mocca, E., Eisfeld, K., Kazepov, Y., Fuzzi, S.,
2021. Public perception of air pollution sources across Europe.
Ambio, 50(6), 1150–1158.
Manohar, K., Brunton, B. W., Kutz, J. N., Brunton, S. L.,
2018. Data-driven sparse sensor placement for reconstruction:
Demonstrating the benefits of exploiting known patterns. IEEE
Control Systems Magazine, 38(3), 63–86.
Mukhtar, I., 2018. With smog season looming, Pakistan shuts
polluting brick kilns. https://www.reuters.com/article/
us-pakistan-climatechange-bricks-idUSKCN1MT0CI.
[Online; accessed 15-April-2022].
Pantavou, K., Lykoudis, S., Psiloglou, B., 2017. Air quality per-
ception of pedestrians in an urban outdoor Mediterranean envir-
onment: A field survey approach. Science of the total environ-
ment, 574, 663–670.
Sun, C., Li, V. O., Lam, J. C., Leslie, I., 2019. Optimal citizen-
centric sensor placement for air quality monitoring: a case
study of city of Cambridge, the United Kingdom. IEEE Access,
7, 47390–47400.
Usama, M., Rahman, A., Khalid, Z., Tahir, M., Uppal, M.,
2022. Urban Air Quality Measurements: A Survey.
Vallero, D., 2014. Fundamentals of air pollution. Academic
press.
Xu, J., Li, J., 2018. The tradeoff between growth and environ-
ment: evidence from China and the United States. Problemy
Ekorozwoju, 13(1).
Yamineva, Y., Romppanen, S., 2017. Is law failing to address
air pollution? Reflections on international and EU develop-
ments. Review of European, comparative & international en-
vironmental law, 26(3), 189–200.
Younis, M., Akkaya, K., 2008. Strategies and techniques for
node placement in wireless sensor networks: A survey. Ad Hoc
Networks, 6(4), 621–655.
Zheng, Y., Liu, F., Hsieh, H.-P., 2013. U-air: When urban air
quality inference meets big data. Proceedings of the 19th ACM
SIGKDD international conference on Knowledge discovery and
data mining, 1436–1444.
Zhou, C., Gao, M., Li, J., Bai, K., Tang, X., Lu, X., Liu, C.,
Wang, Z., Guo, Y., 2022. Optimal Planning of Air Quality-
Monitoring Sites for Better Depiction of PM2. 5 Pollution
across China. ACS Environmental Au.
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLVIII-4/W5-2022
7th International Conference on Smart Data and Smart Cities (SDSC), 19–21 October 2022, Sydney, Australia
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLVIII-4-W5-2022-167-2022 | © Author(s) 2022. CC BY 4.0 License.
173
