Seminary on Montreal accidents such as Methodological courses in GIS at SahelTech

Seminary on Montreal accidents such as Methodological courses in GIS at SahelTech

-

English
14 Pages
Read
Download
Downloading requires you to have access to the YouScribe library
Learn all about the services we offer

Description

This article is a summary of “GIS-based spatial analysis of child pedestrian accidents near primary schools in Montréal, Canada” wrote by Marie-Soleil Cloutier, Jean-Pierre Thouez. In fact, we try disengaging its importance for our country “Mali in Africa” especially all scientific methods used by its experiment authors. We have been not trying to confiscate this article’s properties but we are instead showing GIS’ and Remote Sensing’s utilities to people who are interested about this subject in the other hand we try to make political decision maker paying attention about the high quality of this scientific methods for revolving environmental problems caused by population. Please to cite the original note such as: Cloutier, M., Apparicio, P. & Thouez, J. (2007)-GIS-based spatial analysis of child pedestrian accidents near primary schools in Montréal, Canada, Applied GIS, 3(4): 1-18.
COULIBALY A.B, Student in Geography Information Systems and Management of Natural Resources at SahelTech from 2010 to 2012 under the late Professor KONATE D. responsibility, President of SahelTech in Mali.

Subjects

Informations

Published by
Published 09 May 2013
Reads 35
Language English
Report a problem

REPUBLIC OF MALI
A People-A Purpose-A Faith

Scientia, Virtus, Labore
www.stech.edu.ml




Applied GIS
Theme of seminary


GIS-based spatial analysis of child pedestrian accidents near primary
schools in Montréal, Canada

Wrote by
Marie-Soleil Cloutier, Jean-Pierre Thouez
Geography Department,
Université de Montréal,
Montreal, Canada
ms.cloutier@umontreal.ca
And
Philippe Apparicio
INRS UCS,
Montreal, Canada
Summarized and presented by
Amadou Bina Coulibaly
abina@stech.edu.ml
Managed by
Pr Dialla Konate
dkonate@stech.edu.ml


th thFrom April, 6 to May 7 2012
Bamako/Mali
Contents
I) Abstract, objectives and methodology of the survey
II) Problems expressed in this study: what’s advice for our country Mali?
III) Scientific methods and variables used in this survey: Do all of these variables
have a sense in Mali?
IV) Plan of the survey
V) Role of images in the survey
VI) How this study can be used to improve the management of the environment,
of the human and natural resources?
VII) Conclusions of this survey
VIII) Can we lead the same survey in Mali?
IX) Arguments from this survey permitting to convince a Malian’s authority of
the interest of GIS
X) Comment about GIS’ utility
Conclusion
Reference













1) Abstract, objectives and methodology of the survey
Abstract:
This study is about the accidents occurred on pedestrian in Montreal. It demonstrates
also despite some measures taken that these accidents concerned particularly child between 5
to 14 years. First, they try disengaging the factor of risk by taking elementary school
environment where children are very fluently even most weekday. Second, they try making
suggestion by integrating local environment into the spatial analysis of child pedestrian to
reduce the degree of this problem. And Then socio-economic and environment data into a
geographic information system in order to perform a geographically weighted regression was
done too in this study. So the results that they got, demonstrated that the average network
distance separating accident and closest school is less than 500 meters, thereby confirming a
relationship of proximity between these two locations. These results also demonstrate the
relevance of adding a spatial dimension to the regression model by suggesting that prevention
initiatives should take into account the particular nature of each neighborhood so that more
relevant risk factors can be targeted.
Objectives:
This study wants to be a contribution for integrating the elementary public schools
environment into the evaluation of child-pedestrian accidents. So that means the study period
is limited to the different school calendars (weekdays of the school years - September to June)
and that the proximity between accidents and schools is considered through the construction
of „catchment areas‟.
Methodology:
The first work of scientific research is and stays the methodological research. Indeed,
it consisted to collect as possible as one can necessary documents which treat the theme. Here
for collecting data it was the same case. Thereby, its authors start analyzing some documents
or article as: “Using a geographic information system to understand child pedestrian injury”
viewed by Braddock, M and al. in American Journal of Public Health in 1994; “Child
pedestrian injury in an urban setting: descriptive epidemiology” led by DiMaggio C. and M.
Durkin in Academic Emergency Medicine (2002) and as “The effects of area deprivation on
the incidence of child and adult pedestrian casualties in England” wrote by Graham, D and al.
appeared in Accident Analysis & Prevention in 2005. Thus, these articles such as special or
general books permit their having the first impression even the first ideas about accidents
occurred on pedestrian roads before completing those with data collected on the study area. In
the target to collect, to treat and to analyze data for doing the best prevention, they choose three types of treatments such as point pattern analysis to describe the spatial distribution of
child pedestrian accidents in Montréal; a multiple regression model to explain globally the
number of accidents; and a geographically weighted regression model to show spatial
variations in the relation between the number of accidents and selected explanatory variables.
This methodology is as follow:
 Point pattern analysis
GIS (Geography Information System) and Remote Sensing allow us to analyze data by
using some software as ArcGIS, Arcview, Idrisi and ENVI. Thereby, we can make different
kind of maps because each geographical phenomenon has it particularity whose we have to
notice and to represent on the map with an appropriate symbol.
For example: to represent area on the map we need a polygon (2Dimensions), for road
we have to choose polyline (1Dimension) and finally the point (0Dimension) to symbolize
village… this fact explains why they represent accidents distributions on map with the
point(0Dimension). Furthermore, point pattern analysis was used in this survey as method in
order to detect departures from spatial randomness within the distribution of accidents. Even,
several classical measures were used to describe the accidents sites distribution among which
one can mention: density mapping, nearest neighbour index and standard distance and
standard deviational ellipse, which is a useful way to graphically represent dispersion of point
on the map.
 Regression model and geographically weighted regression(GWR)
This one was the second method used to evaluate the link between selected variables
and frequency of accidents in this study. Indeed, the regression model was replaced by the
GWR because it is not always appropriate with spatial data and it does not take into account
spatial autocorrelation among the dependent and independent variables and it can‟t capture
the non-stationary component of the relationship. Thus the GWR proposed by Fotheringham
and al. in 2002 can help to overcome these drawbacks. Additionally, both of these methods
are based on mathematics calculus.
2) Problems expressed in this study: what’s advice for our country Mali?
In this survey there is question about accident occurred on the pedestrian in Montreal.
So According to the latest World Health Organization injury report, road traffic accidents are
the primary cause of mortality and morbidity in the 5 to 14 year-old age group within North
America particularly in Montréal, Canada. And according to police reports, between 193 and
228 accidents occurred each year, including thirteen deaths and 152 seriously injured victims
(Society of the automotive insurance of Quebec, 2004). Thus, in this study, they want to analyze this situation by doing spatial analysis with GIS‟ softwares. That means to disengage
the spatial interaction existing among accident locations and risk factors. Spatial analysis is
helpful and it is also recommended in this kind of work because it allows a broader scope of
study by visualizing accidents and its surrounding environment on maps and by analyzing
spatially in target to detect locations of high risk and make real decision to reduce it.
Besides, that‟s a very capital survey for Mali. So, view the high number of accidents
occurred in this country which concerns every class of age and all lanes, it is paramount to
make map of accident in order to notice and monitor the area where risk factors is high by
doing spatial analysis in GIS. Another importance of this survey for Mali is the fact that it
allows decision maker being on the alert for intervening quickly. This fact diminishes loss in
human life and economic wasting too.
3) Scientific methods and variables used in this survey: Do all of these variables
have a sense in Mali?
Firstly, Scientifics researches about general and special articles treating the similar
theme were done. Secondly, three types of treatments in GIS (Geography Information
System) such as point pattern analysis, regression model and a geographically weighted
regression model were used in this study. This fact was applied on data collected on the
survey area that means “Island of Montréal, Canada‟s second largest city”. Thirdly, because
of the increasing number of the population on this Island more than 1.8 million people they
founded their investigation on three sources of data: information on schools; police reports on
child pedestrian accidents; characteristics of the built environment around schools. This
method was used in order to model child pedestrian accident risk around primary schools.
Indeed, information on primary schools or School dataset proceeds from the School
Taxation Management Council of Montreal (CGTSIM in French) and includes the current
address, the enrolment for the year 1999-2000 (September to June) and the linguistic
affiliation of each school. Schools that were not open throughout the study period (1994-
1999) are excluded, together with specialized ones, for a total of 331 schools included in the
study (93.8 per cent of the original dataset).
As for police reports on child pedestrian accidents, it is from the Quebec Public
Automobile Insurance Society (SAAQ in French) and it‟s on child-related pedestrian
accidents (0-14 years old) for the period 1995-1999. Furthermore, SAAQ data sets are the
only ones that bring together all police accident reports filled on site by police officers.
Indeed, these reports include information on the victims (age, sex, severity of injuries) and on
the event itself (date, time, location). This dataset includes events involving any contact on the street between a pedestrian and a car that leads to any type of injury for the pedestrian.
Preliminary work was done to allow for a focus on schools: children below the age of five are
excluded (318 accidents) as well as all summertime accidents (223), all undated accidents
(36), and all those occurring during weekends or national holidays (311 plus 74). This led to
the selection of 1335 accidents (58 per cent of the original dataset).
Finally, the Built environment dataset was added to identify socio-economic and
environmental contexts as major risk factors for child pedestrian accidents. Indeed, in this
study four different elements were computed for each school in order to characterize the built
environment: social deprivation, street network density, major road density (as a proxy for
higher traffic) and land-use diversity (entropy index).This latter accomplish the method of
data collection and it came from the 2001 population census at the dissemination area level
and from the City of Montréal Geomatic Division (2004).
We remark here the reliability of the data, because all data included in this study as
Scientifics methods came from primary government sources which create, manage and update
them regularly. And then it proceed from the CGTSIM too which updates the school list every
year and the five different lists (for each year of the study period) were used to find which
school should be kept in our database. Finally, it also came from the SAAQ that fill in the
accident database every year according to police and coroner reports. All of these data were
used in this survey with a plenty of agreement.
After collecting data there was a need to prepare them before doing what there is
question for aiming the target fixed in this study. For that, they procedure as follow: School
and accident mapping, Defining schools‟ „catchment areas‟, Defining the school environment:
(Child population, Social deprivation index, Land use diversity, and Traffic), Defining other
school variables.
Therefore, 331 schools and 1231 accidents were mapped. So you notice that they used
a technical method as defining schools catchment areas in order to precise the results by using
global image, analysis functions such as Costallocation and EucAllocation. This fact allows
us in GIS to determine every accident location for each of these schools in order to let‟s
knowing where accident and its risk factors are important. The school environment was also
definite for preparing data. So it concerned population characteristics, social and demographic
characteristics, land use diversity and traffic or street network. Here we have to remark this
manner to treat data is the essential method used in Geography Information System (SIG).
Indeed, it permits us to disengage density of population in the study area and the factors risks
furthermore, the population at risk too. That means the relationship between occupation and population movements for example: street network and pedestrians. In Montreal, Canada,
there are two kinds of schools (English and French schools). This fact gets an influence on the
mode of transportation of children. That‟s why they considerate it very much in data
preparation. After doing that they make eight variables or classes as follow:
1) Number of accidents per school zone as the dependent variable: This variable was
done in order to know the reasons of each accident in different schools.
2) School language:
- English speaking school: In fact 85(25%) of primary schools are English
speaking. So we have to signalize that all of these schools are so far from
students who are conducted by car. That means few children walk to school
thus less accidents.
- French speaking school concern the stay with more accidents because it is
closest to students.
3) School enrolment: about all schools.
4) Proportion of children: Between 05 to 14 year-old.
5) Social deprivation index: This variable was used to verify the persistence of the
relationship illustrated in previous work and to update it with actual accident
database.
6) Road network density: This variable was the support on which the catchment area
was created.
7) Main road density: This one served the base to definite catchment area too.
8) Entropy index: This latter was created land use classification to measure the land
occupational diversity of the school areas.
All of these variables get a sense for our country, Mali. In fact, we have been assisted
amount of accidents occurred on different part of this country every days. Thus, in this State
particularly in the Capital area, Bamako, there should be a map of accidents with more
explanations about its risk factors in order to fall down the rate of death because an accident.
Finally, these Scientifics methods used above permit us to analyze relationship between
people‟s activities and pedestrians.
4) Plan of the survey
This survey turns around the following points:
Abstract:
1. Introduction
2. Data collection 2.1 School dataset
2.2 Child pedestrian accidents dataset
2.3 Built environment dataset
3. Data preparation
3.1 School and accident mapping
3.2 Defining schools‟ „catchment areas‟
3.3 Defining the school environment
3.3.1 Child population
3.3.2 Social deprivation index
3.3.3 Land use diversity
3.3.4 Traffic
3.4 Defining other school variables
4. Methodology
4.1 Point pattern analysis
4.2 Regression model and geographically weighted regression (GWR)
5. Results
5.1 Descriptive statistics
5.2 Spatial distribution of accidents
5.3 Relationship between accident locations and schools
5.4 Global multivariate regression
5.5 Geographically weighted regression
6. Discussion
7. Conclusion
Acknowledgements
References
5) Role of images in the survey
Pictures are more important in study. Firstly, they constitute a Scientifics support of
theorics methods. In this study for example they plays the same idea. In fact, Census
metropolitan area of Montréal represents the study area on the map. So we can exactly see
which part of world especially of Canada this survey is about. Cf. figure on the following
page:

Over, we were talking to catchment area; there are also pictures whose we can see the
catchment point. For accident pattern distribution which is below we notice on map where
accident occurred fluently and when it is very high or not. That means to let us knowing its
density and to pay attention about these areas. See also the picture below:

Thus, with this kind of map we should be able to pay attention when we should be
driving to the town or walking over round. Most of the time, the decision maker can easily
decide to make the buffer point around accident area in order to avoid a large amount of death
caused by accident.
6) How this study can be used to improve the management of the environment, of
the human and natural resources? We know this study is about child pedestrian accident in Montreal, Canada. It can be
use as an advice or reference by every country which decides to improve the rate of socio-
economic development. In fact, with the population growth in the city, accident is one of daily
problem that people are confronted. Thereby, the research of solution to this problem spends
though some Scientifics methods as spatial analyst used in this study. This fact allows us to
pay attention where accidents are very fluently and to analyze its risk factors. We can also
make a map of pedestrian accident with this study in order to help decision maker about urban
management.
7) Conclusions of this survey
At the end of this study, we can say that their attained the objective assigned to this
survey. So it was to integrate the school environment into the analysis of child pedestrian
accidents and then to disengage its risk factors. For that, they made their point of view on
traffic lights, school crossing guard, children‟s behavior and the mode of transportation to
school as risk of accidents occurred on pedestrian in Montreal. They suggest also physical
barriers for pedestrians as factors to reduce levels of insecurity among children. Finally, we
have to notice this study opens dialogue for another parameter of urban management whose
we can also use GIS‟ methods to prevent and to resolve this in the future.
8) Can we lead the same survey in Mali?
Yes, we can lead the same survey although the Scientifics methods used inside and
according to our realities. In fact, with GIS‟ softwares such as ArcGIS, MapInfo etc. we get
possibility to do spatial analyst of accident and to disengage its risk factors in our country. For
example it possible to relieve accidents points in town with Global Position System (GPS)
and then to definite Buffer area. This map permits us to pay attention on accident point even
to avoid it. That‟s very economics and sanitary.
9) Arguments from this survey permitting to convince a Malian’s authority of the
interest of GIS
This study teaches us that the demographic growth generates the urban development.
This phenomenon constitutes a reel problem for infrastructures already implanted such as
traffics which must also follow this rhythm. It concerns all countries in the world of which the
research of solution implies everybody. One of these matters developed in this survey is
accident occurred on pedestrian in Island of Montreal. It concerned particularly child between
5 and 14 years. Indeed, their authors treat this theme by using some Scientifics methods such
as Point pattern analysis, Regression model and geographically weighted regression (GWR)