We are going to present to you an exemple of traitement : "Calculation and mapping dwelling tax rate within Quito" . Quito is the ecuador capital and a city of 1,5 million inhabitants. Population growth is fairely fast more than four percent by year. This rate make problems when you look at the housing growth possibility in a phisical constrains context with raeged relieve. How can be preserved the historical heritage of the central city zone ? How can be preserved the the  peripherical green ring Where and how can be distributed roads utilities, shops, activities, sawerage utilities? Dwelling tax is source of financial income for the municipalty but also a regulation tool for settlements. A map of tax rate by square meter can be if it fit accurently to the politic choice of  the municipality. To make such map the municipality services can use the informations of computer Quito atlas data base but also the IRD software : Savane.

Savane will allowed them to obtein not only a unique result but a set of different simulations according to the weight given to each parameters.

Within a few minutes time we will present the principles of such a simulation. Of coarse we want to present a method and by no means we want to take the place of municipality autority. So all these result have no operational value.

The data are organised in the data base in geographical layer called "relation", or layer. This non hierarchical approach favors spatial relations, space being the only logical link between the differents relations.

With functions "cart" and "tracer" we have a quick geographical look of the total teritory of the data base; the 7000 blocks of census map are immediatly displayed on the sceen. We can said we have display the base map in classical cartography.
We are going to define a small space in the center of the town we will use this space to hour simulation. Among the nine possibility we have,
we are going to chose "definiton direct".

This method consit in delineating a square on the screen. Ce cadre de travail ainsi défini, la fonction « wind » permet de connaître les caractéristiques précise du cadre, c’est à dire les coordonnées géographique du point de référence, la largeur de la fenêtre et la taille du pixel.
When defining the frame of our work by the square, the function "wind"  gives the precise carateristics of the square meaning geographic coodinates of of the reference point, the width of the square and pixel size.

 

For a real duty, the choice of criterian and the procedure may be queit a long work that one should not negleted. here our choice is quiet quick. Our tax rate will take into account three main sets of variables or criterias. A criteria of quality and confort housing, a second criterian is the access to different facilities whether public or private, at last a criterian of quality or contraint of environment, slope and vegetation.

 

For each criteria we calculate an index by a succession of specific query : respectivly

A query about housing  data

An other set of query about infrastructure datas

A third one about environment data

Taxation rate is then calculated by lineary combining of three indexes.

 
 

Let us first create the index about housing confort.

. Where are the data we should use to compute this index ? i.e. room number, sanitary ???, house construction quality ? Data base contien a layer number 15. This relation, which name is INEC from the name of Institut that collect housing data for the 1982 census.

 
 

We see that the data base include several geographical layers named relations. For the first one, INEC, about dwelling,  data was collected at block level, each block is caracterise by 54 attributes. Whit the "list" menu we can see the data base thematic parameters. We can find the room number as the 22th attribute.

The dwelleing number , by block, the tird attribute, And sanitary utilities, from the 30 th to 33 th attributes.

 
 

the building materials , from 23th. to27. th attributes.

This "relation" is a set of areas. To make a good map we need some specific spatial entities : They to not be too big, if they are too big they might aggregate some very different states. So we think that "block" is the refence that will be used by taxation office. So better used this geographic units to map the final result. This geographic unit choice is a very important step because it's going to lead all the further calculation.

For each block, whatever was the first geographic refence of data, we will choice to compute by simple linear combining : (X=aAttribut_1+b Attribut_2..) The"cris" function allowed to make this type of treatment, for exemple dividing room number by dwelling number we create a new attribut : the average of room number by dwelling.

 

Result will be the weighted sum of three parameters : - room number by dwelling, - dwelling with sanitary utilities - numbers and built up  quality. The peculiar class of without sanitary utilities will give zero value and will not stop the treatment.

 

If we want to make some more complexe analisis we can link to SAS software. Now we did data analisis and classification. Then we had draw and ware the result map of the combination  that we had called : "comfort index" with button "stat", "???", "draw".  The comfort index is growing up from light colors to dark colors. This map is just one possible display in five quantils of intermediate result. Data comfort index still in unbroken form, in "real" values.

Now, we calculate the index : "infrastructure access possibilities". Numerosus attributes can be take in account, we are just going to select three of them :   school capacity, sewer linkage and urban roads proximity without using other attributes like "public garden", "ealth utilities", "phone line", "water supply", "market place", etc ..... thought they are in the data base. Remember : in Savane the geographic frame of each data, by spot, line or zone, have to correspond to the collect unit. We a going to see how to cross several layers which have different geographic caracteristics.

First step : school capacity analyse. This data, about school, is include in the "education" layer. The geographic characteristic of this relation is "spot", each scool is located by one point.

 

Rather than using a distance model (distance model image) we use the school official areas map.

 

But we have a big obstacle : there is no attribut number of school area for the blocks. How can we manage ? But we can use the official school areas

 

The process is : summing the numbers of children of all the schools present within each official school area.

Each official school area contien now a new attribut : "student capacity". Second step : each block get the value of the official school area within it is located. We have a map by block of school access index.

Then sewage linkage :

We get the sewer system in a linear type "relation". We need of a binary information : the block is linked or not. The "mask" function allowed to make a buffer zone of all spaces whithin a one hundred buffer around primary sewer network.

The "restriction in the mask" function applied to the INEC "relation" within thre mask we just made give us the map of all the blocks linked to the primary sewer network. (in purple on the map).

 

Urban road proximity index We get the main urban road network in a linear type "relation".

Final index we want have to give the mean distance to the main road of each point of the space. From the "main road relation" we compute this distance for each point. Each point of the resulting map present a value , in meters, that is the distance to a main road. The distance,in meters,is growing up from dark colors to light colors.

We have compute the three index :" school capacity", "sewer accessibility", "road proximity". That enable us to calculate the whole "infrastuture index" by block. For each block the new attribut will be weighted sum of weighting factors of each three indexes : Result= a*School + b*Serwer +c*Road .       Arbitrary we chosed a=1,3; b=1.5 and c=10. Savane allow to create a new layer logicaly combining several data from very different geographical type. We see also that is the user himself who decides.

If we change weight of value we can obtien results very different (here a=10, b=2.5 and c=1.3). Using macro commands we can get a great number of different simulations.

The third criteria is about environment.

For this third criteria we will going to use raster treatments. First slope in urban fabric zone. (The building part of the city located on on height slope is very costly to maintein for the Quito municipality). One "topographic layer" include the contour lines (every 20 meters) and another one the altitude point (accurancy 10 centimeters) for building areas at the mapping date. With these both data we elaborate a DEM in which each pixel will have a calculed altitude value (in meters)

We can display it in classes of altitude, like beside.

Or in  shading off mode ..........................
To create an attribute "mean slope" by block, we first, compute the value for each pixel. Second we calculate  a average value by block. We transfered topographic data in administrative units.

It may be useful to weight the tax rate by the green space presence. We can get a green space rate map from the Thematic Mapper satellite image. These data is pertinant as its spectral caracterises allowed to extrac with accurancy the green cover. We get to create 9 type classification. Savane create a new layer about vegetation with 5 of these classes.

Then we can calculate and assign an average index for each block. In same way that previous process we assigned to administrative units, data originaly available in raster type.

Environmental factors map is obtain by linear combining of two criterias : slope and vegetation. At least, the final result is a tax rate map including the three chosen criterias. It worth to paid attention to the powerful sofware tool used but also to the leading role of treatment choices especialy the choices of differents parameters. This necessary rigour is an avantage because allowed discussions the rate with autorities and public presentation of this process is a good way to explain the rate elaboration to urban dwellers. Because the final product it's not only one map but numbers and rules.