![](bg2.png)
a city manager. From the dashboard, a traffic man-
ager can pick a road to be analyzed and, with a cer-
tain detail, visualize the traffic profile and its patterns
with the possibility of comparing it from different in-
stances. Also, the environment of the city can be cor-
related with traffic information in a way of detecting
how traffic affects the pollution in the nearby areas,
providing indications of anomalies. From the results
presented in the dashboard, it is detected bad traffic
zones on the streets in different hours of the day, visu-
alize that a certain event also affects the traffic when
comparing the same street in another date, and ob-
serve that higher levels of traffic have direct impact
on the air quality.
The paper is organized as follows. Section 2
presents the related work in this area and some
dashboards implemented in smart cities. Section 3
presents the system architecture and their compo-
nents, and the algorithms and procedures made when
handling information. Section 4 exposes the final so-
lution with use cases for analysis. Section 5 concludes
the results of the paper and presents the future work.
2 RELATED WORK
2.1 Traffic Analysis Approach
TrailMarker (Honda, 2016) is an example of a work
where different approaches are used to analyze real
data sets of vehicular sensor data to find outliers, ef-
fectively finding typical patterns and points of vari-
ation on the road. This work relates to the present
work by having similar objectives of finding road pat-
terns, but is more centralized on the driver side pat-
terns where acceleration and speed are used to predict
the type of driver (inexperienced, aggressive, careful).
Erdogan (Erdogan and et al., 2008) propose a geo-
graphical information system that has the objective of
assisting the analysis of road accidents based on acci-
dent reports. In the system, the accident locations are
geo-referenced into the highways so it is easier to de-
tect the areas with high rate accidents, in a way to take
precautionary measure and improve the safety in criti-
cal areas. Another important topic is the study of pol-
lution levels related with the road traffic. In an anal-
ysis and evaluation made on the main urban roads in
Beijing (Li et al., 2002), it was used an in place noise
sensors along the road for multiple sampling points
in order to identify the factors influencing the noise
levels. From their results, where more than 3000 ve-
hicles pass every hour, above their designed capacity,
the noise levels exceeded the national standard dur-
ing daytime hours. Other factors were also pointed,
the road width, surface texture and the types of vehi-
cles in the traffic composition. A proposed model for
the evaluation of urban traffic congestion using buses
(Carli and et al., 2015) uses GPS data from bus on-
board units to probe the position and time trace and
road segments, to associate the GPS traces and calcu-
late the traffic congestion. It offers a map, graphical
and tabular representation of traffic patterns. A work
(Bacon et al., 2011) that uses real time bus data makes
an integration of data from multiple sources (buses,
cameras, OpenStreetMap), and calculates the journey
times depending of the time of the day based on bus
real time positioning. A model proposed (Kerminen
and et al., 2015) uses bus location history data to an-
alyze the traffic fluency and the time spent on driv-
ing, on bus stops and traffic lights, and the frequent
delay areas of the city during the vehicle journey to
understand the causes of delays on bus lines. In (Zhu
et al., 2012) it is studied the road traffic conditions us-
ing data collected from taxis where it was probed the
location and driving speeds during the movement. It
uses an algorithm based on compressive sensing that
achieves low error with missing data.
2.2 City Dashboards
Many cities are starting to make changes in their intel-
ligence in different areas by adding sensors and pro-
viding information through dashboards. The city of
Dublin contains a mobility dashboard
1
where the traf-
fic can be seen in a global view in the main roads, but
also parking lot locations. Our proposed work can ex-
tend the information presented by Dublin dashboard
by analyzing the roads with greater detail to improve
anomaly detection, with information in all streets.
The dashboard of London
2
presents many city
variables from points of interest, public transportation
(metro status), environment (weather and pollution
data) and cameras (street view) are viewed in real-
time. Although there is a lot of real time information,
this is only being captured and presented for a gen-
eral consumer, lacking on a more detailed analysis on
traffic congestion and pollution data.
3 PROPOSED SYSTEM AND
ARCHITECTURE
We propose a full stack solution as a system that inte-
grates information from the city to a city level dash-
board as exposed in figure 1.
1
http://www.dublindashboard.ie/pages/DublinTravel
2
http://citydashboard.org/london/
VEHITS 2018 - 4th International Conference on Vehicle Technology and Intelligent Transport Systems
388