1. Improving Traffic in Oulu
Team Blue
Jarkko Vatjus-Anttila, Matti Pouke, Nervo Verdezoto, Matteo Picozzi
2. Introduction
● We decided to pick the vehicular traffic data and Wifi/BT access
data as the basis of our teamwork.
● The key questions for the work:
● Who: City traffic-related decision-makers and planners.
● What: A tool, which is able to show temporal information
about both vehicular and pedestrian traffic.
● Where: City of Oulu, perhaps nation wide audience
● Why: To help understand temporal traffic phenomena better,
and to help decision makers to detect problematic patterns,
which should be taken into account in city planning.
● How: With a combination of data mining and web-based
viewing tool, which allow to inspect temporal traffic data as
charts, as well as fit into the city map.
6. Interviews
● We went and visited the center of traffic control
authorities, asked a few questions about the current
solutions, and finally showed off a demo or ours
● We got three different views on the topic:
● Traffic management employees
● Local traffic police
● City planner
7. How does the old system look like?
Black&White map, with green intersections
(color has no meaning here) How to get intersection history
data:
1) RMB select one intersection
2) Select ”properties”
3) Select time/date, and export
4) Import to Excel
5) Draw a graph!
6) Goto 1) if other intersections are needed
History data exists, but no statistical
analysis nor graphical representation of it.
8. Inteview Results:
Generic comments:
Oulu traffic Oulu traffic police officer Oulu city planning
management:
● They have realtime ● Real-time calculation and ● Most important thing for
information of the overall prediction of the traffic city planners is the
traffic system waves would be very pedestrian traffic
valuable information for estimation, because a lot of
● But they do not have any Police group management designs and decisions are
statistics. They have only made based on this.
excel exports and manual ● An ability to place units
processing ahead where the masses ● However, this info is not at
are traveling, would be the moment available real-
● Traffic light re-programming AWESOME time, and hence this new
is slow, since data not system would be
easily available ● Pedestrian traffic found AWESOME!
very valuable and
● Graphical, realtime, interesting. Especially ● Nowadays, they put
statistical analysis of the when delivered realtime. someone to stay at one
traffic is AWESOME, which intersection and count
they have been requesting ● Overall, they were amount of people
for some time already impressed of a demo done manually...
in less than a week. It was
● Alas, it is too expensive, much better than any of the
they have been told :| recent ”demos” that they
have seen.
9. Inteview Results:
Improvement ideas
Oulu traffic Oulu traffic police officer Oulu city planning
management:
● A series of traffic posts and ● Statistical traffic volume ● Pedestrian movements
their cumulative traffic over combined to accident realtime
time is very valuable for database with location
traffic light reprogrammers. information is a wanted ● Ability to offer optimal
solution places for companies to
● Programming of ”green have their outlets, based
waves” especially ● Statistical estimation of on the amount of
problematic at the traffic for areas where there pedestrian traffic in
moment. are no sensors would be different places.
very useful.
● These exists a static model
of oulu with adjusted traffic ● Separation of traffic
capacity per road. It could be between light and heavy
combined with the statistical vehicles is desired.
model of the esitmated traffic
data ● Reading back sensor data
from Nokia Maps, would
● The sensor data that was enlarge the sensor
given to us does not match coverage.
the capacity nor organization
of the intersections. This
would be needed for ”final”
version.
10. Interview Results:
Typical use-cases
● Traffic planning: re-programming of traffic lights
● Now: export number of slices of history data for given
intersections. Import them to excel. Calculate statistics. Calculate
new optimal timings. Re-program the lights. They can do 20 lights
at a time now.
● With our tool: They can select the time frame of interest. They
can see staistics immediately for day/night periods separately and
evaluate the reprogramming need immediately.
● Police: moving of units in the field
● Now: Midsummer comes and historycally, police knows when,
how and to where people are leaving the city. If the guess fails,
re-placement of units in the field needs to be done.
● With our tool: From realtime data traffic vectors are calculated
and they see all the time the ”traffic wave” as how it exists. They
can immediately react to the change of mass. The most important
factor is time.
11. Conclusions
● We apply information visualisation techniques to get
insights about traffic in Oulu.
● Based on the interview feedback we validated our
hypothesis and ideas.
● During the project we learned a lot about data mining,
data processing and information visualisation
techniques.