Dynamic Exposure Visualization of Air Quality Data with

Augmented Reality

Sylvain Renault

, Ingo Feldmann

, Lieven Raes

, Jurgen Silence

and Oliver Schreer

Fraunhofer Heinrich-Hertz-Institut, Einsteinufer 37, Berlin, Germany

Digitaal Vlaanderen, Havenlaan 88, 1000 Brussels Belgium

Keywords: Augmented Reality, Air Quality Data, Environmental Data, Visualization, Citizen Science.

Abstract: Increased awareness regarding air pollution and environmental conditions is more relevant than ever. Notably,

there is already widespread availability of do-it-yourself (DIY) environmental and air quality sensors across

Europe. These data are accessible to the public through various web interfaces, providing insights into current

and past environmental conditions across different regions of Europe. Augmented Reality (AR) stands out as

a promising technology to enable citizens to monitor environmental conditions in their communities and

comprehend their own impact, thus motivating behavioural changes. Nevertheless, effectively visualizing

real-time environmental data in the 3D AR space remains a challenge. Innovative visualization techniques are

necessary to present environmental data in a clear and engaging manner. In this paper, we introduce a

framework, a visualization concept, and a prototype AR application capable of providing a visual overlay of

environmental information such as air quality or traffic data. These concepts stem from the European project

H2020 COMPAIR, which involved users throughout the development process. The application will undergo

evaluation in various pilot cities and regions and will be publicly available via app stores by mid-2024.

1 INTRODUCTION

Due to the availability of affordable do-it-yourself

(DIY) environmental and air quality sensors, there is

already a good coverage of sensor data across Europe.

Similar trends are observed in other regions such as

India (EOS, 2022). These sensors are typically linked

to platforms where users can access real-time and

historical environmental data. One such established

platform is Sensor.Community, which operates a

globally-contributed sensor network generating Open

Environmental Data (Sensor.Community, 2023).

Various web interfaces facilitate data visualization

and exploration, making the information accessible to

the public. Figure 1 showcases a 3D representation of

nitrogen dioxide distribution in London

(MappingLondon, 2023). Utilizing Augmented

Reality (AR) for understanding environmental

conditions offers a more intuitive approach. With just

a smartphone or tablet, users gain access to a visual

https://orcid.org/0000-0003-2015-2573

https://orcid.org/0000-0002-2595-8683

https://orcid.org/0000-0001-9284-462X

https://orcid.org/0000-0001-6900-8287

overlay displaying environmental data like air quality

or traffic conditions. This advanced visualization in

3D AR enables direct monitoring of air pollution

levels in the surrounding area. Increased awareness of

local environmental conditions can influence

behaviour, promoting alternatives such as biking or

using public transport. Such awareness may also

drive changes in local government policies, such as

Figure 1: NO2 distribution in 3D (MappingLondon, 2023).

120

Renault, S., Feldmann, I., Raes, L., Silence, J. and Schreer, O.

Dynamic Exposure Visualization of Air Quality Data with Augmented Reality.

DOI: 10.5220/0012586700003696

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2024), pages 120-127

ISBN: 978-989-758-694-1; ISSN: 2184-500X

implementing measures to tackle high air pollution

levels near schools.

While prior attempts at developing similar AR

applications exist, many are no longer accessible on

major app stores. This paper introduces a new concept

and outlines the implementation of an AR application

designed for mobile devices. It can visualize real-time

environmental data from various Open Data

platforms. The scientific contribution lies in diverse

visualization techniques customized for distributed

spatial data within the augmented environment.

Developed within the European research project

H2020 COMPAIR (H2020 COMPAIR, 2023),

extensive user involvement throughout the

development process ensured thorough examination

of innovative visualization concepts in four pilot

cities. The resulting modules and assets were

integrated into an Air Quality (AQ) Framework.

2 STATE OF THE ART

Without a doubt, AR holds significant promise for

enhancing education and learning, a fact supported by

various surveys (Vargas 2020, Patel 2020,

Khanchandani 2021). However, the exploration of

visualizing environmental data within AR applications

remains relatively unexplored. Few published AR

applications concentrate on presenting air pollution

data for mobile devices, but with limited geographical

reach. To our knowledge, there is currently no AR

application accessible that offers augmented visuali-

zation of traffic data sourced from public open datasets.

Torres and Campbell presented an AR app (Torres

and Campbell, 2019) visualising main pollutants

from the World Air Quality Index Project (WAQI,

2019). In 2019, it was available on the App Store, but

not anymore. Mathews et al. presented AiR, which

was available for Android devices (Mathews et al.,

2021) (see Figure 2). In both applications, pollutants

are displayed with flying dots and depending on the

severity of the pollution, the number of objects

increases. Additional functionalities are provided

such as access and visualization of historical data.

The

AiR app

was designed for residents of India,

utilizing data from the country's 222 monitoring

stations, which, considering India's size, forms a

relatively sparse network. Developers said they might

create an iOS version, but this plan never

materialised. Furthermore, the last update on Google

Play was in 2020, and the last comment in 2021.

Despite a promising start (Google Play reviews,

publication in IEEE), the team behind AiR

discontinued the app. In the same year, the New York

Figure 2: AiR app (Mathews et al., 2021).

Times added an AR feature to its app that visualised air

pollution based on location (New York Times AR App,

2019). However, this app does not visualize real-time

sensor data, but visualizes microsized pollution

particles that were floating around on the worst day of

the year at your current location. The Helsinki Air

Quality AR app provides Air Quality data from Finnish

Meteorological Institute, but only for the city of

Helsinki. The last update was offered in2022. In

Campana & Dominguez, (2021), a scientific prototype

of a particulate matter visualization in AR is presented,

but the app did not yet become public.

In this paper, we present a first prototype that is

developed in the COMPAIR project. This Dynamic

Exposure Visualization App (DEVA) accesses various

servers that provide a wide range of sensor data

distributed across Europe. The primary challenge lies

in presenting a manageable amount of information in a

clear and easily understandable manner. The app

provides an alternative view of sensor data as depicted

in a 3D visualisation environment of an Urban Digital

Twin. Hence, we present a novel visualization concept

for the representation of environmental data in 3D AR

space. DEVA can run on any mobile device, either

tablet or smartphone, which has the necessary tracking

and AR visualization capabilities.

3 THE PROPOSED SYSTEM

The implementation of the application and the AQ

framework is based on Unity 2022 LTS i.e. using the

common user interface package uGUI, which allows

the rendering of the user interface in 3D space. We

also used the AR Foundation Framework, a unified

workflow to build robust apps for various AR devices

(Unity AR Foundation, 2023). To transpose geo-

localisation information between the GPS-based

sensor positions and the world coordinate system of

Dynamic Exposure Visualization of Air Quality Data with Augmented Reality

121

the AR scenery, the framework uses an external GPS

asset (Unity AR+GPS Location, 2023).

A system overview of DEVA is depicted in Figure

3. The application is able to receive sensor data from

two different sources: from project internal sensors

and DIY sensors as proposed by the

Sensor.Community (Sensor Kits, 2023). Both group

of sensors send their values via a PHP respectively

RESTful interface to the two servers supported by

COMPAIR. DEVA is then able to request all current

as well as past sensor data via the dedicated API. As

the interface definition is defined in OpenAPI

(OpenAPI, 2023), it is easy to adapt the interfaces to

the requirements of server and client.

The AQ Framework in Figure 3 shows the

processing workflow within DEVA. The

Communication module receives the sensor data via

REST requests. The data are then transformed in an

OGC compliant data structure. The Open Geospatial

Consortium (OGC) offers a standardised definition

for (i) geospatial content and location-based services;

(ii) sensor web and Internet of Things; and (iii) GIS

data processing and data sharing, as defined in the

OGC SensorThings API Part 1 (OGC, 2022).

The Pipeline Management then processes the

sensor values and their metadata and passes them to

the different Unity Engine 2D and 3D graphic

modules. A number of visualization methods are

implemented to turn the sensor data into 3D

geometries. One important issue of the app is the data

accessibility and the access to AQ sensor data from

various servers or services over the internet.

3.1 Environmental Data Servers

The COMPAIR project distributed DIY air quality

and traffic data sensors to citizens in four different

cities in Europe in order to capture a variety of

environmental data. The air quality sensors from

SODAQ (SODAQ, 2024) provide data such as CO2

(Carbon Dioxide), NO2 (Nitrogen Dioxide), BC

(Black Carbon), PM1, PM2.5 and PM10 (Particulate

Matter of size 1, 2.5 and 10 micrometres), humidity

and temperature. The traffic sensor (Telraam, 2024)

offers vehicle-counting information for individual

classes such as cars, trucks or bicycles. The

measurements of those sensors are collected and

managed in a centralized Data Manager hosted by

COMPAIR. Additionally, the app accesses the

European Sensor.Community open data server, to

include additional AQ data. Both services are auto

configured by the app, so that a user can immediately

start with exploring sensor data.

Once connected to the servers, the app requests

the data in series, from each of them. Those requests

are made for the surrounding environment of the geo-

localized user, the close area where the user interacts

with the data. When the user moves through the

streets, the app regularly updates the sensor data. To

reduce the transfer payloads, data requests can be

restricted to specific sensor types by activating a

filter.

3.2 Geo-Localisation

A proper AR visualization requires an accurate

localization and orientation of the user and its device

in space. Hence, the AQ framework contains three

geo-localization modules (GeoPose, 2023), using the

Global Navigation Satellite System (GNSS): (i) data

request depending on the GPS position of the user;

(ii) the geo-pose of the mobile device; and (iii) the

geo-localized air pollutants for visualization in the

AR/3D space.

The app observes the activity of the user while

moving to present actual 3D data. It requests new

surrounding sensor data in intervals as follows. The

current GPS position is retrieved from the device as

the user location in the world. Then, it continuously

measures the current distance from this position to the

Figure 3: System overview.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

122

centre of the current loaded dataset. Finally, if the

distance exceeds the half of the view perimeter e.g.

500m for a maximum distance set to 1km (see Figure

4), the module starts the requests of new sensor data

from the Data Manager and Sensor.Community. This

approach limits the amount of data to be transmitted.

Once new data are downloaded, they are

transposed in the internal OGC structure, injected into

the pipeline via the container system and visualized

by the AR Content Renderer. Old containers and their

3D geometries are removed before new ones are

visualized. The Pipeline Management System takes

care of the correct 3D positioning of data as described

in the next section. The data update may require one

to many seconds, depending on the perimeter size and

the amount of loaded data.

4 AR VISUALIZATION

The main purpose of the app is to present

environmental real-time sensor data in AR to the

citizens. Past attempts on AR visualisation of

pollution information relied on quite simple

approaches, e.g. visualizing pollutants via flying balls

or rocks. The major scientific challenge is to develop

and investigate different approaches that suit best for

a meaningful representation of a large amount of 3D

data in the 3D AR space. This led to three different

approaches following the Near’n-Far concept (see

Sec. 4.2). This concept optimizes the placement of

data that are far away from the user. In collaboration

with the project partners, we are able to perform an

end users evaluation of the novel visualization

techniques. Once, sensor data are received and

accessible in the OGC format, the framework

enriches the data with graphical objects for the

rendering step. Therefore, three processes are

initiated: (i) the creation of the asset visuals from the

current selected design elements; (ii) setup of the

asset attributes corresponding to the sensor type; and

Figure 4: Update technique while the user is moving.

(iii) the placement of the objects in the AR virtual

environment in the native coordinate system used by

the render engine.

4.1 Data Visualization

The app offers various sensor visuals rendered in the

AR view. All designs are created as a Unity prefab

consisting of assets and scripts. The user can freely

select a design appropriate to the user’s level or taste.

Three classes of sensor value representations in the

AR space are implemented (see Figure 5 and 6):

 a simple version uses 3D primitive meshes like

spheres, cubes or squares. They are very easy to

render because of the low complexity of the

polygon models and their material attributes;

 a more complex representation combines simple

3D meshes, 2D pictograms and GUI elements,

like text fields and button canvas for values,

sensor types and meta-data;

 a cloud visualization relies on particle animation

exploiting the capabilities of the Graphics

Processing Unit (GPU) of the mobile device.

One or many clouds are created depending on

the density of the sensor data, by using a

DBSCAN clustering algorithm (Density-Based

Spatial Clustering of Applications with Noise).

This algorithm observes points in space that are

close together and derives related clusters of

arbitrary shape. Per cluster, the cloud

visualization then creates one particle animation

system based on cloud/fog tiled images. Clouds

are animated by the render engine and offer a

special dynamic to the AR view, because of its

3D volume.

Once a visual was instantiated, the properties of

the prefab are set-up according the sensor type. Thus,

not all property fields are used for all objects, e.g. a

simple visual does not have a text field for the name

of the sensor or metadata. The following attributes are

Figure 5: Various visuals for the same values (simple,

complex, complex with gamification).

Dynamic Exposure Visualization of Air Quality Data with Augmented Reality

123

Figure 6: Cloud visualization, one colour for all sensor

types vs. multi-coloured (simulation).

implemented to setup the visual of the graphical

elements:

 Colour: Each pollutant corresponds to a specific

colour and is also used in the legend. The

gradient of this colour illustrates the current

value's environmental impact.

 Text: Sensor values and units are displayed in a

UI label field, utilizing a Unity text canvas. In

some cases, the sensor type and brief description

are shown in an additional field.

 States: In state mode, four icons indicate (i)

active sensor, (ii) calibrated data, (iii) values

within an acceptable range, and (iv) private data.

Colours change between red-yellow-green to

indicate states.

 Size (optional): Prefab size varies based on value

impact (small for normal, large for high values).

 Transparency (optional): Data transparency is

used to render distant data with decreasing

opacity.

4.2 Near’n-Far Concept

In the presented AR visualization framework, we

propose a new concept to rearrange geo-located data

in the 3D AR space. The main purpose is to minimise

occlusions in dense areas with many buildings, trees

etc. and to allow the user to capture easily the

distribution of 3D data in the AR space. Hence, we

developed a new method considering the

visualisation of data in the near and far distance

separately. This render method helps avoiding some

perspective and visualisation conflicts, so that the

user is not anymore confronted with incoherent

situations and can get a better understanding of the

information in the AR space. The following issues are

addressed with the Near’n-Far concept:

 Overlap: The geo-pose of sensor devices usually

does not contain the altitude. Hence, all sensor

values are located in the same height and the

overlap of data increases by the density and

values cannot be distinguished from each other.

 Perspective: The framework requests data for a

perimeter surrounding the user, freely set by the

user e.g. 1km. However, in some areas, only a

few measurements are available due to sparse

distribution of sensors. Therefore, enlarging the

perimeter is an option to get more data in the AR

scene. The problem then is that data become

very small because its visibility decreases with

distance.

 Occlusion: When travelling in a city, a user is

walking or driving between buildings or trees

that create potential occlusion. As the app does

not use a 3D map (e.g. Mapbox, 2023) or object

detection, objects do not occlude sensor visuals

and data remain always visible. Hence, the

perspective parallax of the real and the virtual

objects contradicts each other in the AR view.

The Near’n-Far solution is similar to a technique

named “proximity”, where the user can interact

adequately with a device at different distances (De la

Barré et al., 2009). A proximity system places objects

of interest in a main area with the possibility of

interacting with them. The other objects are placed in

an area beside, or in the background, and are

visualized in a minimized fashion. The Near’n-Far

concept relies on a double perimeter system: the inner

perimeter is dedicated to near data and the outer

perimeter to far data (see Figure 7). The initial data

container in the pipeline is split into two new ones:

(i) a near data container manages the data inside the

inner perimeter and uses the same visuals and

positions as originally;

(ii) a far data container for data with a distance

bigger than the inner perimeter using a simple

visual, e.g. small spheres, with new calculated

xyz positions and sizes.

While activating the Near’n-Far mode, the user can

change the inner perimeter in real-time and the

pipeline manager will update all active containers

accordingly. The centre point of the Near’n-Far

object is always the geo-location of the user. While

moving, the system updates all data containers. In all

modes, only the geometric formulas transforming the

positions of far containers differ. We implemented

three different Near’n-Far methods in the render

pipeline:

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

124

Figure 7: Near’n-Far concept.

Theatre: The first mode is an amphi-theatre like

presentation, where data are arranged in different

heights depending on their distance to the user (see

Figure 8). We reduced the intensity of the AR view in

the screen shot for better visibility of the data. Closer

data are rendered on the same level as in the near area.

Beyond, data rise up to the heaven. Hence, the user

perceives the distance of the data by its altitude and

size. This results in fewer conflicts of the far data with

the real world.

Dome: The second mode is a dome-like presentation

as depicted in Figure 9. The dome is an invisible half-

sphere around the user. The formula is based on the

one from the theatre, added with a projection of the

far data onto the dome centred to the user’s location.

Closer data are arranged in the lower part of the half-

sphere, while distant data are located higher up on the

half-sphere. All data laying on the sphere has the

same best optimal size. Depending on the

environment, the radius of the sphere should be

adapted: in a dense city location, a smaller sphere

should be more suitable (e.g. inner perimeter to 50-

100m).

Ring: A third option is a virtual ring that surrounds

the user (see Figure 10). The geometrical formula is

based on the one from the Dome with a projection of

the sensor locations onto the ring. The idea is that the

ring is close enough to the user (5m), so the user can

easily select sensor data by clicking them. A rotation

mode with a compass will facilitate the user to access

information around him, without the need to turn the

body or device. In Table 1, we compare the

capabilities of the three different visualization

methods according to the aforementioned issues. The

dome representation respects all three aspects from a

conceptual point of view.

Figure 8: Theatre: Top-left: Concept; Bottom-left:

Authoring and Bottom right: Screenshot.

Figure 9: Dome: Top: Concept; Bottom left: Authoring and

Bottom right: Screenshot.

Figure 10: Ring: Top: Concept; Bottom left: Authoring and

Bottom right: Screenshot.

Dynamic Exposure Visualization of Air Quality Data with Augmented Reality

125

However, the Near’n-Far concepts presented, are not

capable to visualize too many or too densely

distributed data. In such case, it is important that the

user filters the values by sensor types. Still, the

dynamic re-ordering of data via the Near’n-Far

approach offers the user a better visual experience, if

sensor data collide with the real world.

Table 1: Comparison of different visualization methods.

Method Overla

Pers

ective Occlusion

Theatre YES NO YES

Dome YES YES YES

Rin

NO YES YES

4.3 User Interface

In our app, the graphic engine Unity renders the full

screen. The 2D user interface is an overlay of the

3D/AR viewport and does not require any operation

system (OS) specific UI libraries. Our framework

offers a number of assets and reusable Unity

components allowing to adapt and extend the UI as

desired. The app supports a simple and an expert

mode, which is a result from our in-depth user

evaluation especially with school kids. We identified

that the UI should be used with minimal technical

skills, no prior knowledge, self-explanatory and

intuitive in use. By designing the 2D screen, we

respected standards such as responsive design, white

spaces, organic elements up to demonstrative

gamification of the UI, such as animations and

pictograms. The UI adapts automatically to the

orientation of the device and its screen resolution,

especially to support both, mobile phones and bigger

sized tablets (see Figure 11).

Figure 11: Main elements of the GUI: Left: Hamburger

menu and Right: Toolbox

4.4 Trip Recorder

To encourage citizen participation, we integrated a

feature known as the trip recorder (see Figure 12).

Upon activation, this tool records and saves the

current GPS position at various time intervals. Once

the trip concludes, the position and time data are

transmitted to the Data Manager for subsequent

analysis. This tool serves as the input for another

project-developed application called the Dynamic

Exposure Visualization Dashboard (DEV-D),

available for access at (DEV-D, 2024). This separate

application monitors and visualizes air quality along

the recorded trip.

Figure 12: Trip recorder.

5 USE CASES

The main objective of DEVA is to offer the wider

public the possibility to get awareness about the

environmental conditions in their local

neighbourhood. In the best case, this may change

behaviour and will have impact on political decisions

on the local level, e.g. implement low emission zones

or change rules for traffic and develop mobility plans.

In the context of the COMPARE project, a fully

citizen science (CS) based approach is performed to

involve citizens in many aspects of the project such

as installation of DIY sensors in the local

surrounding, participating in workshops, or

contribute to user testing and evaluation of various

tools developed in the project.

The close cooperation with citizens is

implemented in five different pilot cities in four

different European countries. The pilots are designed

around the following topics:

 investigating air pollution over different periods

in time (weekend vs. working days),

 investigating environmental conditions in

school streets,

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

126

 measuring effects before and after certain

mobility plans in a neighbourhood.

Especially for the development of the presented AQ

Augmented Reality app, the CS involvement is

crucial and very helpful. We were able to receive

immediate feedback from end users in every stage of

the development, starting with the requirements, the

first prototype until the final application.

6 CONCLUSION

The presented Augmented Reality app offers for the

first time a real-time visualization of air quality and

traffic data based on publicly available environmental

DIY sensor data. The overall framework of DEVA

followed a flexible and dynamic software concept,

which can easily be extended with additional

functionality according to the user needs and the

availability of new sensor data. For example,

inclusion of gamification elements could be deployed

in future versions, to increase the attractiveness and

joy of the app, especially for the younger audience.

The main contribution of this paper is the proposition

of three different kind of visualizations for 3D data in

the 3D AR space. Thanks to the H2020 project

COMPAIR, the development of the different

prototypes was continuously accompanied by project

partners. Hence, the UI design, the necessary

functionalities are a result of in-depth testing and

evaluation thanks to the CS-based nature of the

project. The app will be made available to the public

by middle of 2024 on app stores.

ACKNOWLEDGEMENTS

The research presented in this paper has received

financial support from the European Union’s Horizon

2020 Program COMPAIR under grant agreement no.

101036563.

REFERENCES

Campana F.E. and Dominguez F.X. (2021). Proposal of a

Particulate Matter Measurement Device and an

Augmented Reality Visualization App as an

Educational Tool. 12th Int. Conf. on Education

Technology and Computers (ICETC '20). Assoc. for

Computing Machinery, New York, NY, USA, 6–11.

De la Barré, R., Chojecki, P., Leiner, U., Mühlbach, L.,

Ruschin, D., (2009), Touchless Interaction-Novel

Chances and Challenges, 13th Int. Conf. on Human-

Computer Interaction. Part II: Novel Interaction

Methods and Techniques, Berlin, Germany, Springer

Verlag, Heidelberg, Germany, pp. 161–169, July 2009,

doi: 10.1007/978-3-642-02577-8_18

DEV-D – Dynamic Exposure Visualization Dashboard

(2024): https://monitoring.wecompair.eu/ dashboards/

dev-d

EOS (2022): https://eos.org/articles/indian-cities-invest-in-

low-cost-air-quality-sensors

GeoPose (2023): https://www.geopose.org/

H2020 COMPAIR (2023): https://www.wecompair.eu/

Helsinki Air Quality AR (2021): https://play.google.com/

store/apps/details?id=net.playsign.hip

Khanchandani, K., Shah, M., Shah, K., & Panchal, V.

(2021). A Review on Augmented Reality and Virtual

Reality in Education, Int. Research Journal of

Engineering and Technology (IRJET), Vol.8, Issue 2,

February 2021, https://doi.org/10.1186/s13173019-

0084-8.

Mapbox Web Services API (2023), Mapbox Inc.,

https://docs.mapbox.com/api/

Mathews, N.S., Chimalakonda, S., & Jain, S. (2021). AiR:

An Augmented Reality Application for Visualizing Air

Pollution. 2021 IEEE Visualization Conference (VIS),

146-150.

MappingLondon (2023): https://mappinglondon.co.uk/20

18/3d-map-of-nitrogen-dioxide-pollution/

New York Times AR App: https://vrscout.com/news/nyt-

app-ar-air-pollution/

OGC (2022): OGC SensorThings API Part 1: Sensing

Version 1.1, https://docs.ogc.org/is/18-088/18-088.html

OpenAPI (2023): https://www.openapis.org/

Patel, S., Panchotiya, B., Patel, A., & Aishwariya

Budharani, S.R. (2020). A Survey: Virtual, Augmented

and Mixed Reality in Education. Int. Research Journal

of Engineering and Technology, Vol. 9, Issue 5, May

2020.

Sensor.Community (2023): https://sensor.community/en/

Sensor Kits (2023): https://sensor.community/en/sensors/

SODAQ (2024): https://sodaq.com/

Telraam (2024): https://telraam.net/

Torres, N.G., Campbell, P.E. (2019). Aire: visualize air

quality. In ACM SIGGRAPH 2019 Appy Hour

(SIGGRAPH '19). Assoc. for Computing Machinery,

New York, NY, USA, Article 1, 1–2.

https://doi.org/10.1145/3305365.3329869

Unity AR+GPS Location, Dev. by Daniel Fortes, 2023:

https://assetstore.unity.com/packages/tools/integration/

ar-gps-location-134882

Unity AR Foundation Framework, 2023: https://unity.com/

unity/features/arfoundation

Vargas, J.C.G., Fabregat, R., Carrillo-Ramos, A., Jové, T.

(2020). Survey: Using Augmented Reality to Improve

Learning Motivation in Cultural Heritage Studies.

Applied Sciences. Appl. Sci. 2020, 10, 897.

WAQI – The World Air Quality Index (2019)

https://aqicn.org/here/

WEForum (2023): https://www.weforum.org/agenda/20

16/10/air-pollution-facts/

Dynamic Exposure Visualization of Air Quality Data with Augmented Reality

127