Using Raster Based Solutions to Identify Spatial Economic Agglomerations

class RS_Densitate { //member data // object that signifies no argument protected object miss = Type.Missing; // the cell size protected double cels; // the name of thematic layer for which the raster image will be generate protected string numel = null; // the thematic layer which will generate the raster image protected ILayer l_dr = null; // the working space protected IWorkspace ws = null; // extreme colours (initial and final colour) protected IRgbColor ci, cf; // the generate raster protected IRaster ries = null; // the map document protected IMxDocument doc; // the map protected IMap harta; // rendering the raster image protected IRasterRenderer rsrend = null; // the factor for adjustment values, depending on the measure unit protected int fmum = 0; // the constructor that takes as parameter the map document, // extracts the actual map and // build two objects, for storing and manipulating colours public RS_Densitate(IMxDocument phdoc) { ci = new RgbColorClass(); cf = new RgbColorClass(); doc = phdoc; harta = doc.FocusMap; } // the property used to set / get the cell size public double Dim_cel { get { return cels; } set { cels = value; } } // the property used to set / get spatial data, by the thematic layer name public string Data_sp { get { return numel; } set { numel = value;} } // the property used to set / get the workspace public IWorkspace Sursa { get { return ws; } set { ws = value; } } // setting colours, based on red, green and blue values public void Set_culoare_start(byte vr, byte vg, byte vb) { Informatica Economică vol. 17, no. 1/2013 135 ci.Red = vr; ci.Green = vg; ci.Blue = vb; } public void Set_culoare_end(byte vr, byte vg, byte vb) { cf.Red = vr; cf.Green = vg; cf.Blue = vb; } // pure virtual method used to generate the raster protected abstract void Generare_Raster(); // the default display of raster public void Afisare_Implicita() { // to generate raster – the method aims // setting IRaster type variable Generare_Raster(); // a raster type layer is created, by using IRaster type variable IRasterLayer rl = new RasterLayerClass(); rl.CreateFromRaster(ries); // add raster layer to the map harta.AddLayer(rl); // raster layer moves to the last position, to assure non overlapping other // spatial elements from the map harta.MoveLayer(rl, harta.LayerCount 1); // the TOC (Table of Content) is updated doc.Update


Introduction
There are a lot of scientific papers supporting the idea that in the presence of agglomeration economies, the potential for growth is increasing in the level of economic activity [1].The huge amount of spatial data generated by GIS (Geographical Information Systems) expansion, the increasing number of geographic informatics applications available, the computerization of a large amount of information sources, and the availability of digital map has increased the opportunity and need for the usage of methods and techniques for spatial data analysis and integration with economic data, for both research and applied purposes.[3] This paper proposes a GIS based solution to automate the identification of agglomerations in space, by determining the density of spatial elements, function on their physical locations.The proposed solutions suppose to use software components developed to automate and to reuse the same behavior, for a large set of applications, developed for many domains and economic applications.The software components could be used in scenarios based manner.Scenarios developed in a certain field, can be reused in the same domain, extending or improving them, or in other scientific fields, by adapting them to new requirements.The process of reuse is dependent on the complexity of the information presented, describing the use of functions implemented readiness of users in areas that have been developed for these scenarios.[5] The proposed solutions use raster concept, as support to generate the level of density of spatial distributed economic items, i.e. the density of companies performing in one economic field, as they are distributed in space and the distribution of communication means like roads, in space, based on their densities.A raster consists of a matrix of cells, each containing a value representing information, such as: temperature or the presence of one economic phenomenon in a specific place.Rasters are digital aerial photographs, imagery from satellites, digital pictures, or even scanned maps.Data stored in a raster format represents realworld phenomena, like: thematic or discrete data, representing features such as land-use or soils data and continuous data, representing phenomena such as temperature, elevation or spectral data such as satellite images and aerial photographs.Within a GIS, the uses of raster data fall under the following main categories:  raster as a basemap by using orthophotographs -it is a background display for other feature layers.This kind of raster is used to display underneath other layers, and provide the map users with the confidence that map layers are spatially aligned and represent real objects, together with representing of additional information.There are three main sources of raster basemaps: from aerial photography, satellite imagery, and scanned maps;

1
 raster as a surface map -are well suited rasters for representing data that changes continuously across a surface.They provide an effective method of storing the continuity as a surface.They also provide a regularly spaced representation of surfaces.Elevation values measured from the earth's surface are the most common application of surface maps, but other values, such as rainfall, temperature, concentration, and population density, can also define surfaces that can be spatially analyzed;  raster as a thematic map -are raster representing thematic data, which can be derived from analyzing other data.Thematic maps can result from geoprocessing operations, which combine data from various sources, such as vector, raster, and terrain data.For example, you can process data through a geoprocessing model to create a raster dataset that maps suitability for a specific activity;  raster as attributes of a feature -rasters used as attributes of a feature may be digital photographs, scanned documents, or scanned drawings related to a geographic object or location [6].The main advantages of storing the spatial data as a raster are: the simple data structurebecause a matrix of cells with values representing a coordinate and sometimes linked to an attribute table; a powerful format for ad-vanced spatial and statistical analysis; the ability to represent continuous surfaces and to perform surface analysis; the ability to uniformly store points, lines, polygons, and surfaces; and the ability to perform fast overlays with complex datasets.[6] 2 Programmatic Solutions To accomplish the tasks of this paper, solutions based on the raster processing are used.To develop the software components the ArcObjects framework is used.ArcObjects is a framework developed on COM (component) type technology, used to build ArcGIS software, developed by ESRI (Environmental Systems Research Institute) [4].By using a framework to develop software applications, the advantage that allows developers to reuse code and also to facilitate the building of new extensions is supported.In order to use the raster based symbology, that allow us to reflect the spatial data density in the territorial area, a hierarchy of classes has been developed.The hierarchy of classes enables the use of specialized coclasses for generating the symbology, based on spatial data types like point and polyline.The coclasses are used to generate new objects [3]. Figure 1 presents the developed of classes hierarchy.It is an inheritance relationship between classes, in the sense that coclasses inherit attributes and methods of the base abstract class.Raster type symbology can be generated, from spatial data having specific types like: line or point.RS_Densitate class has been defined as an abstract class, because it cannot be directly instantiate.The Generare_Raster method used to generate the raster has been described as a pure virtual method because it cannot be concretized inside the abstract class, but it makes sense to be defined the concrete classes derived from it.Generare_Raster method aims to generate raster, by using the input parameters, some are common for both types of spatial data, and others are specific for each type of data.Once obtained, the raster can be appropriate symbolized, so that the areas from geographic studied are able to be displayed as suggestive as possible, in terms of density of spatial data generated.Because the symbology applies to the resulting raster, the specific methods used to generate the symbology were grouped into parent class (the abstract one).For this purpose three methods have been defined:  Afisare_Implicita used to display the raster by default, i.e. by using colors from black to white; the black color is associated with the lowest density areas, and white color is associated with the highest density areas, intermediate density areas are coloured in shades of gray;  Afisare_Continua used to display the raster by using continuous tone of colors, the user provides, for this purpose, the start color that will match the lower density areas and the last color has to be associated with the highest density areas; intermediate areas will be colored with transitional tones, that make the transition from initial to final colors;  Afisare_Discreta used discrete color tones to display the raster.The user can define the number of groups used to divide the area, according to the calculated densities.Thus, a number of intervals of values have to be set out, for the density values, and a color will be assigned for each group (range) of values to view these areas.In this case, the extreme colors will be provided, and the intermediate groups will be then filled with transition colors.To have a high degree of generality, for the hierarchy of classes, a constructor has been defined.The constructor has the goal to create the connection with the document map, which is generated based on raster symbology.The input parameters are provided in the form of properties, which are of read-write type.To display the density of spatial data like a raster, a raster with only one level has to be developed.The matrix stores the density values, in the required points, so that on displaying, the numeric values will be scaled on a color ramp.Thus, the points turn differently depending on their values.To have a better accuracy in representing the densities, the raster has to contain an array of real values (double).Get_minmax method aims to calculate the minimum and maximum values of the raster matrix.Simbologie_Raster_Interval method aims to symbolize the raster, so that the pixels to be colored according with the values membership to a range of values.In this way, the user can specify the number of groups, namely the number of intervals that have to be generated, and the pixels from each group will be separately colored.This view allows the user to mesh the raster, in terms of colors, so that the diversity of matrix of pixels values will be displayed, by using fewer colors.This method has to surprise important changes in density within the geographical area studied.The method gets the number of groups (z), which is the number of intervals of values that has to be generated, and vsup parameter given by a symbolic constant which indicate the measure unit used in the labeling.Suprafata enumerative constant consists in The code execution generates and allows displaying a raster, as presented in figure 3. The color space is scaled and it is associated with pixels having the values between the two extremes.In figure 3 we can observe that areas with a higher density of roads are intensely colored with red color, while other areas are colored with a more diluted red, so that no roads in an area, causes it to be colored in white.The density is calculated as the sum of the length of roads, per unit area.Being a map with geographical coordinates of projection type, the considered measure units could be meters (by default) or kilometers.ity of proposed solutions, in the paper, we are trying to identify the cluster type economic agglomerations, in the pharmaceutical industry of Romania.To accomplish this task we have to identify the agglomerations of companies acting in pharmaceutical industry, located in a geographical proximity as the first step, and the second step supposes to identify the geographical areas with a good agglomeration of transport facilities (roads).In the figure 5, the geographical distribution of companies performing in pharmaceutical industry of Romania is figured, by using points.On the same map, the layer with the available roads has been represented.

Fig. 5. Spatial agglomerations in pharmaceutical industry of Romania
By using the software components developed, as they are presented in this paper in the previous sections, the density of companies (points) and the density of roads available to link them (lines) have been identified, in each geographical area.The results are figured on the same map, and are presented in figure 5.A strong economic agglomeration of companies acting in pharmaceutical industry could be identified in the South part of Romania, in Bucharest-Ilfov-Prahova area.Even there are a lot of agglomerations of companies performing in this field, in various geographical areas, because there is not good enough transport infrastructure available, to link them, we can conclude that there are not good conditions to identify a successful economic cluster in pharmaceutical field, in these regions.

Conclusions
To display in GIS the economic phenomena, namely the statistical data, both symbologies based on spatial data and raster symbology can be used.Symbologies associated with spatial data are limited to their location, while raster based symbologies facilitate the definition of areas that are strictly related to the phenomenon of spatiality.The raster based solutions facilitate to integrate the spatial and economic data and to perform analysis like the spatial distribution of economic issues in space.The software components presented in the paper automate the processing of both economic and spatial data and generate the appropriate maps containing both types of data.

Fig. 1 .
Fig. 1.The hierarchy of classes used to display spatial data density in space Figure 2 illustrates this structure for an image having m x n resolution.

the name of thematic layer for which the raster image will be generate protected string numel = null; // the thematic layer which will generate the raster image protected ILayer l_dr = null
RS_Densitate has the following content: