Development and Integration of an Offline and Open Source

Alternative Mapping Solution

Antoine Abélard

Spécinov, Angers, France

Keywords: DotSpatial, Open Source GIS, C#, .NET, OpenStreetMap.

Abstract: Nowadays, more and more business applications integrate Geographical Information System (GIS) to use

and display data at the right place and the right time. However, this integration could face two problems.

The first one is the integration cost of a commercial GIS and the second is cartographic data access. Indeed,

the web map services used to provide cartographic background are charged depending on queries, so the

cost of these services can increase quickly and the bill is not mastered. The second problem has occurred

with the end of Microsoft MapPoint, so, this brings up the problem to find an offline mapping solution that

can be integrated with business applications. There is a need to fill this gap and answer to these two main

problems. This paper presents an offline and open source alternative which allow integration with business

applications, a real business case of a truck tour management application has initiated this work.

1 INTRODUCTION

GIS and mapping technologies have grown in

popularity in recent years, especially since the

success of Google Maps, Bing Maps and Here

Maps, all available on the web. However, these GIS

platforms do not meet the requirements of some

business applications such as the cost and the offline

mapping.

The main issue of this study is about a business

application of truck tour management called Biotour,

it is a desktop application used by SARIA Industries

(SARIA Group, 2015) and developed by Spécinov

(Spécinov, 2015). SARIA Industries is specialized in

collecting organic waste and biomass conversion.

The group operates over 127 sites, in 12 European

countries. SARIA Industries has four areas of

expertise: Rendering, Food industry, Oleochemicals

and Energy.

Spécinov is an IT services company localized

next to Angers in France. Working in partnership

with a university research laboratory, Spécinov

developed 11 years ago this tour management

application named Biotour, used by SARIA

Industries for his rendering activity in 20 sites in

France. This activity involves the collection,

transport and treatment of animal waste. The

software is designed to calculate the best collection

routes to optimize the human and material resources.

This application was developed in VB and C# on

.NET platform and used Microsoft MapPoint

(Microsoft, 1999) for his cartographic rendering. But

since MapPoint announce the end of his support, the

sustainability of Biotour is threatened. Moreover,

MapPoint was not built to manage multiple data

layers and as important volumes of business data. So

there were stability issues which stop the application

or slowing it. In fact, MapPoint was the right choice

11 years ago due to the offline solution which is a

key point for this application.

So, today, the goal is to build a replacement

solution which is offline and with a limited cost. The

chosen solution will have to implement the same

functionality as MapPoint, to display data with the

same rendering and to be more stable.

To build the alternative solution we have to find

the two main features of a mapping solution: a GIS

platform and a data provider. In the MapPoint

solution, both are integrated.

The present work is divided into 8 sections. We

present the data provider OpenStreetMap in section

2. Section 3 introduces the GIS platform DotSpatial.

The application architecture is given in section 4.

Section 5 details the management of data layers in

the solution both concerning the cartographic layers

and the business layers. In section 6 we describe the

management of interactions. Afterwards, section 7

Abélard, A.

Development and Integration of an Ofﬂine and Open Source Alternative Mapping Solution.

In Proceedings of the 2nd International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2016), pages 139-144

ISBN: 978-989-758-188-5

139

shows the work done on integration and

optimization. We conclude with the results and the

feedback of this new solution in section 8.

2 OpenStreetMap: THE DATA

PROVIDER

OpenStreetMap (OpenStreetMap, 2004) is a

crowdsourcing project to create a free map of the

world. This project was launched 11 years ago and

provide now a high-quality world map coverage

(Girres, 2010) (Haklay, 2010). OpenStreetMap data

are under Open Database License (ODbL). This

license agreement intended to allow users to freely

share, modify, and use this dataset while maintaining

this same freedom for others, provided that the

original source and author(s) are credited.

The main advantage of OpenStreetMap data is

that they are free and can be downloaded for an

offline use. Indeed, there are no many data providers

that allow to download data for an offline use. In our

study case we need France cartographic data, there

are only two data providers for an offline use,

OpenStreetMap and IGN (French National

Geographical Institute). However, the reference

pricing of IGN for the France data is €160,000, so

the cost constraint is important. Therefore,

OpenStreetMap was chosen as the data provider of

the application.

For our use with the GIS platform which has

been chosen, OpenStreetMap data will not be used

in their default format but in the shapefile format. It

is possible to convert OpenStreetMap data default

format, osm, in a format compatible with GIS

platform with some tools. But, for simplicity

reasons, we get OpenStreetMap data directly in

shapefile format from Geofabrik (Geofabrik, 2007)

server. Indeed, Geofabrik does a daily extraction and

split data by region, which allow us to manage

basemap by region. Moreover, shapefiles provided

by Geofabrik are also split by data type (roads,

railways, places, …). So it allows to manage

independently each data layer depending on data

type.

3 DotSpatial: THE GIS

PLATFORM

DotSpatial is an open source GIS library written for

.NET 4 which has been developed by members from

the Geospatial Software Lab at Idaho State

University, Idaho Falls, Idaho, USA (Idaho State

University, 2015), as well as an international open

source community (Ames, 2007). It allows

developers to add spatial data, analysis and mapping

functionality into their applications, it provides a

map control for .NET and several GIS capabilities.

The DotSpatial architecture follows two

paradigms. The first is a separation of the user

interface and business logic. The second is to use a

larger number of modules to increase code

reusability, and allow access to certain parts of the

library without requiring a whole. So, DotSpatial is

divided into 17 DLL and just some of them are used

to build a GIS platform needed by the business

application (Cao, 2012). The reasons for choosing

this library has been its ability to manage large

volumes of data, its flexibility and mainly the size of

the community mobilized around this project. This

open source alternative for GIS allows the

development of specific software (Sameen, 2014);

(Wang, 2014).

4 APPLICATION

ARCHITECTURE

Figure 1 illustrates the Biotour application

architecture.

Symbology

editor

BioTour

.dspx

C#

Interface

library

DotSpatial

library

shapefiles

SQLServer

Figure 1: Application architecture.

The Biotour application being developed in

Visual Basic .NET, the choice is made to develop a

C# interface library that will access the DotSpatial

library. It creates an abstraction layer between the

business application and the library that allows to

manage all function calls in the interface library and

to work at a higher level. For example, when loading

the cartographic background, DotSpatial called a

function from the interface library which then uses

the DotSpatial library to perform this action.

As can be seen from the architecture presented

above, the DotSpatial library uses several elements.

First, to display the cartographic background, a

.dspx file (XML) is used. This file contains the

GISTAM 2016 - 2nd International Conference on Geographical Information Systems Theory, Applications and Management

140

description of layers to import and the symbology to

apply to them. This file is editable through a

symbology editor included with the library.

The cartographic data from OpenStreetMap are

stored in shapefiles, they each contain a data layer

(roads, natural features, places...) for a defined area.

Once the data from shapefiles were loaded, the

symbology defined in the dspx file is applied and the

basemap is displayed in the map control. Then,

business data like the Biotour network and the

customer location points are displayed on additional

data layers created from the data stored in the SQL

Server database.

5 MANAGEMENT OF DATA

LAYERS

DotSpatial allows, like any GIS solution, to manage

the data in a system of map layers. Among the layers

used in the map controls, some are part of the

basemap and are static layers. Other layers are

business layers on which the user will be able to act

(selection, drawing, additional data display, etc.).

The map data layers are loaded from shapefiles

while business data layers are loaded from the

existing SQL Server database of Biotour.

Businesslayers

SQLServer

Cartographiclayers

shapefiles

Figure 2: Layer types and their origin.

Next, when we look at the representation of data

on a map, the notion of map projection comes in. In

all online mapping services, such as Google Maps or

Bing Maps, the projection used is the Web Mercator

projection, this projection is also used by MapPoint.

Because of the constraint to remain as close as

possible of MapPoint rendering, this projection is

used in the new solution. However, coordinates of

basemap data are not projected in Web Mercator but

are in WGS84 format (a standard for use in

cartography, geodesy, and navigation including by

GPS), so a projection of this data is required. These

projections are made when the dspx file containing

the legend is loaded. About coordinates of business

data present in the database, they are also in WGS84

format. So, conversion functions for switching from

one projection to another have been implemented to

project coordinates from WGS84 to Web Mercator

and vice versa.

5.1 Cartographic Layers

Data of these layers are collected under the shapefile

format. In this format, each entity has metadata that

includes information on it. For example, for a

segment of roads data file, there is the type, the

direction of movement, etc. To display raw data in

an understandable representation, a symbology must

be applied to each data layer based on metadata.

Symbology operates on the association of a

representation and a filter expression. The filtering

works with simple logical expressions that are

applied to the metadata. When a symbology is

applied to a data layer, if the metadata associated

with an entity corresponds to the case of a filter

expression then the corresponding style will be

applied to that entity. Here are some examples of

filter expressions:

• [TYPE] <> 'primary' AND [TYPE] <>

'secondary'

• [NAME] = 'Rue Victor Hugo' AND

[ONEWAY] = '0'

Data displayed on the map depends on the zoom

level. However, in the DotSpatial library, the

function to manage the appearance of a data layer

depending on zoom level is implemented but does

not manage the appearance depending on metadata.

Thus, the appearance of data is set for the entire

layer and it is therefore not possible to display only a

part. So, this behavior had to be implemented by

modifying the symbology when the user changes the

zoom level.

A basemap which look likes all others basemap

provided by web map services has been

implemented. The symbology is applied both from

the .dspx configuration file for the majority of layers

and from the interface library which allows to

implement a symbology depending on zoom level.

5.2 Business Layers

Data of these layers are loaded from the existing

database of the application. No changes of the data

structure are required to build a map layer from an

existing table. A map object is built from a business

object by assigning his coordinates to the map object

and all others properties to the metadata of the map

object. Then all built map objects are assigned to a

layer that can be displayed with the appropriate

symbology. Inversely, to get a business object from

a selected map object, coordinates and all metadata

of the selected object are used to build a business

object that can be used in a business process.

Development and Integration of an Ofﬂine and Open Source Alternative Mapping Solution

141

In the same way as for data of cartographic

layers, metadata attached is used to manage the

symbology assigned to each entity. For example, the

items loaded in the pick-up points data layer will

contain in metadata the period from which the pick-

up point is waiting for, to have a differentiated

display of points according to this criterion. Unlike

basemap, data of business layers are modifiable and

will take into account the interactions, so it is

possible to draw or select elements in these layers.

Data from the basemap layer are loaded and are

not altered. By cons, data of business layers can be

loaded, modified or deleted. So, add, modify and

delete functions have to be implemented for each

business layer.

6 MANAGEMENT OF

INTERACTIONS

The next step to build a mapping solution is the

management of interactions. Indeed, beyond display

data, a mapping solution has to manage user

interactions with these data.

The first one is the selection. When an item is

selected on the map control, an event is raised and it

must be processed to get the selected item on the

selected data layer.

Then, MapPoint can display a tooltip that

contains information about the clicked point, but the

selected framework does not implement it. So, this

feature has been implemented in the interface library

using the default tooltip of .NET platform.

Another interaction is the position and distance

calculations. The position calculation provides a

geographical position from the cursor position on the

map. For example, it is used in the network design

for positioning a point of a new segment. The

distance calculation provides the distance between

two coordinates, it is particularly used in the

network design to assign a length to a drawn

segment.

The print feature is used in the roadmap printing

feature. This roadmap is printed for each driver, it is

a representation of the tour and pick-up points on the

basemap. This feature has been implemented with

the new solution, a map control is loaded and a

snapshot of the tour loaded on the basemap is

generated and integrated into the roadmap.

The last one is a location search function. The

implemented feature allows a repositioning of the

map at the position of the desired location. To do

this, the user must enter the name of the desired

location, a function search if some of maps objects

have a name which corresponding in their metadata

and return the results. Then, if results are found they

are displayed and the map can be repositioned on a

found location.

7 INTEGRATION AND

OPTIMIZATION

Finally, once the alternative mapping solution was

built, the final step is its integration in the business

application.

Following the integration of the mapping

solution in Biotour, it appeared that the map is

refreshed at every addition or modification of data in

a layer. This behavior leads to waiting time due to a

large number of refreshment called when adding or

editing multiple data following. To optimize this

operation and prevent unnecessary map refreshment,

additional functions were set up in the library to

disable and enable the map events. So before we add

or edit data, events are blocked. When we will

reactivate it after acting on the data, all events will

be queued up and the map is refreshed only once.

This greatly reduces the waiting time of the user

when he acts on the map data.

To avoid loading several times the data of the

basemap the map control used in the application is

unique. But even by doing it this way, loading the

map data is still required at least once. This time is

about 1 to 2 minutes, which may seem long for a

user. To avoid having a visible latency that time will

be masked by loading data in a dedicated thread.

We can see the result of the integration in the 3

following figures. In figure 3 we can see the window

of Biotour network drawing in which the Biotour

network can be modified over the basemap. Figure 4

is the window of manual management of generated

tours which allows to adjust manually the generated

tours. In this figure, we can see that the symbology

applies to the basemap data is not the same as in

figure 3 to have a better readability. Finally, figure 5

shows the window of pick-up points selection before

tours generation.

GISTAM 2016 - 2nd International Conference on Geographical Information Systems Theory, Applications and Management

142

Figure 3: Window of Biotour network drawing.

Figure 4: Window of manual management of generated tours.

Figure 5: Window of pick-up points selection.

Development and Integration of an Ofﬂine and Open Source Alternative Mapping Solution

143

8 CONCLUSIONS

To conclude, an innovative alternative mapping

solution which is offline and open source was built.

It has much more flexibility than MapPoint and

others existing online mapping solutions. In each

screen of Biotour, the implemented solution allows

the user to see only layers that he needs to see.

Further, unlike MapPoint, data of the basemap can

be daily updated from OpenStreetMap to always

having quality data ensuring two features

completeness and correctness. Finally, the new

solution is much more stable than MapPoint so the

user experience is improved.

A key point is that the project takes place in the

case of a real business application, with customer

constraints. The flexibility of the new solution

allowed to implement customer needs that can’t be

implemented before with MapPoint, so the interest

of the map in the application is reinforcing.

Moreover, the application is more stable and offer

important development opportunities. Therefore, its

sustainability is now ensured.

The new version of Biotour using the new

mapping solution has been evaluated and compared

with the previous version which is in production.

The results show that the new solution is better in all

criteria : flexibility, rapidity, stability, ability to

work with large amounts of data and especially

sustainability. Indeed, the new Biotour application

works without MapPoint now, so all the defects of

the application, which are inherent to MapPoint,

were removed.

This new version is now deployed in all Saria

sites and users’ feedback is very good. They all

noticed a significant improvement of stability and

quality of the application. So, beyond the scientific

success to build an innovative mapping solution with

specific constraints, it is also a business success.

These successes show that another way to integrate a

mapping solution is possible.

ACKNOWLEDGEMENTS

This work was carried out thanks to ESEO,

Spécinov and Saria Industries.

REFERENCES

Saria Group (2015) – Specialists in the recycling of animal

and vegetable products and waste products.

http://www.saria.com/

Spécinov (2015) – IT services company.

http://www.specinov.fr/

Microsoft (1999) – Software program created by

Microsoft that allows users to view, edit and integrate

maps. https://www.microsoft.com/mappoint/en-us/

home.aspx

OpenStreetMap (2004) – Collaborative mapping project.

http://www.openstreetmap.org/about

Girres, J. F., & Touya, G. (2010). Quality assessment of

the French OpenStreetMap dataset. Transactions in

GIS, 14(4), 435-459.

Haklay, M. (2010). How good is volunteered geographical

information? A comparative study of OpenStreetMap

and Ordnance Survey datasets. Environment and

Planning B Planning and Design, (37), 682-703.

Geofabrik (2007) – Company providing shapefile

extractions of OpenStreetMap data.

http://www.geofabrik.de/geofabrik/geofabrik.html

DotSpatial (2010) – Open Source Geospatial Framework

in .NET. http://dotspatial.codeplex.com.

Cao, Y., & Ames, D. P. (2012). A strategy for integrating

open source GIS toolboxes for geoprocessing and data

analysis. In Proceedings International Environmental

Modelling & Software Society (iEMSs) 2012

International Congress on Environmental Modelling

and Software, Sixth Biennial Meeting.

Idaho State University (2015) – Members from the

Geospatial Software Lab developed DotSpatial.

http://www.isu.edu/

Ames, D. P., Michaelis, C., & Dunsford, T. (2007).

Introducing the MapWindow GIS project. OSGeo

Journal, 2(1).

Sameen, M. I., Ali, A. A., Wahid, N. O. A., & Kubaisy,

M. A. A. (2014). An Approach to Develop a

Geographic Information Database Using Dot Spatial

Open Source Platform and Google Search Engine. J

Comput Sci Syst Biol, 7, 217-220.

Wang, R., Liu, N., Xu, M., & Kong, X. (2014). Research

on the Open Source GIS Development Oriented to

Marine Oil Spill Application. Journal of Software,

9(1), 116-120.

GISTAM 2016 - 2nd International Conference on Geographical Information Systems Theory, Applications and Management

144