House rent modeling for Khulna City

Abstract

In our country, housing projects are generally instigated according to the zoning regulation. This is the common technique of planning to take a housing scheme by development authority. The factors or criteria’s associated with the sustainable and economically potential planning for housing are often ignored in the contemporary planning practices like zoning. Due to the projection difficulties of rent structure, most of the housing projects do not assure the cost recovery of the projects as well as the projects tend to be up to snuff. House rent is depended on several criteria or facilities surrounded or related with it. General criteria affected the house rent are types of house, better accessibility to city center, the floor space, the environmental condition and the level of utility and services facilities. There are other several criteria that affect house rent. It is tried to explore those criteria in this study.

The main purpose of this study is to develop a linear regression model of house rent. Vector base GIS software ARC info, ARC view and ARC GIS as well as statistical analytical software SPSS has been used to illustrate this spatial decision support model for Khulna City.  With this model, the development authority as well as other public or private organizations can chose their site for housing considering the cost recovery and desired profit.  It has been tried to build a linear regression model that can explain the house rent of Khulna city in the study. With this model one can choose the house according to his ability in one way and other way one can fix the house rent for an existing location. With this model future house rent can also be predict. Therefore this study mainly deals with the spatial data for the decision model development. But in reality, other non-spatial information needs to be considered. So further study should be carried out considering non-spatial information. There have been no attempts to develop house rent model in Khulna City. Taking for instant, this study is intended to pave the way for subsequent studies, development and uses of urban house rent models in Khulna City.

1.1 Background

A large proportion of residents in cities and towns of developed as well as developing countries are tenants. Renting and sharing in the cities of developing countries are still neglected topics. Statistical link between housing tenure and the level of economic development is less than clear. Home ownership is lower in many developed countries than in most of the developing world. As UN-HABITAT has noted: “the fact that many of the world’s richest countries have a large rental sector demonstrates that home ownership levels cannot be taken as a symbol of national prosperity (UNCHS, 2001a:96). UN-HABITAT also concludes that “governments should review their housing   policies and device appropriate strategies for rental housing which remove biases against non owners”. (UNCHS, 1990:6). In Bangladesh on an average 33.66 percent households live in rental housing (BBS, 1990). In Khulna City about 25.66 percent people live in rental housing (BBS, 1994). With the rapid increase of urban population there is a need of rental housing in Khulna city. Rental housing projects should be instigated in such a way that it can recover its cost within a desired period of time. So there need a model of house rent so that any public as well as private organization can initiate their housing project securely as the model shows appropriate house rent on a particular site.    

The term ‘model’ entered the lexicon in the 1960s when the idea of symbolically representing complex system suddenly came in to age. This was as much due to computer reaching the point where large data set could be routinely manipulated as it was to any fundamental shift in our understanding of complex system in science or society (Batty, 2001).  Generally, a model is a simplified representation of the planner’s understanding of the development process in his or her particular planning situation, i.e., a particular planning jurisdiction, over a particular time period, for a particular planning problem (Batty, 2001).  Models are idealized in the sense that they are considerably less complex and subtle than situation being represented; only the most relevant factors and processes are included in the model. By simulating urban behavior, they help the planner anticipate where future problems will arise. They also express a systematic statement of relationships among elements of the problem situation and thereby improve the planner’s understanding of the forces of change and the relative importance of various factors in change process.

Urban planners have been using computer models since the early 1960s when computers become available for public uses. Numerous planning models were developed since then. However, it is not until recently that planning models become more commonly used. The recent development of computer technology such as geographic information system (GIS) has been created new opportunities for model development and use. It provides a framework for integrating spatial and non-spatial information from different format, sources and time period. With these capabilities, it is relatively easier to generate and incorporate various spatial variables in urban models. However, these capabilities have not been fully utilized to date especially in the context of developing countries.

In our country, housing projects are generally instigated according to the zoning regulation. This is the common technique of planning to take a housing scheme by development authority. The factors or criteria’s associated with the sustainable and economically potential planning for housing are often ignored in the contemporary planning practices like zoning. Due to the projection difficulties of rent structure, most of the housing projects do not assure the cost recovery of the projects as well as the projects tend to be up to snuff. House rent is depended on several criteria or facilities surrounded or related with it. General criteria affected the house rent are types of house, accessibility to city center, the floor space, the environmental condition and the level of utility and service facilities. There are other several criteria that affect house rent. It is tried to explore those criteria in this study. The main purpose of this study is to develop a linear regression house rent model. Vector base GIS software ARC info, ARC view and ARC GIS have been used to illustrate this spatial decision support model for Khulna City.  With this model, the development authority as well as other public or private organizations can chose their site for housing considering the cost recovery and desired profit.  There have been no attempts to develop house rent model in Khulna City. Taking for instant, this study is intended to pave the way for subsequent studies, development and uses of urban house rent models in Khulna City.

1.2 Objectives

      To identify the factors / variables that influence or determine the house rent

      To develop a linear model that can explain the house rent for Khulna City.

1.3 Rational of the study

There are three perspective that articulate the need of house rent modeling

1.3.1 UN-HABITAT view

A large proportion of residents in cities and towns of developed as well as developing countries are tenants. Home ownership is lower in many developed countries than in most of the developing world. (UNCHS, 2001a; p.96). In 1989, a meeting of experts organized by UN-HABITAT concluded that governments of developing countries should review their housing   policies and device appropriate strategies for rental housing which remove biases against non owners.

Renting in the cities of developing countries are still neglected topics. This negligence of rental housing has, however, not been wholly uncontested. A number of studies have appeared over the last two decades. During the 1990s, a number of researcher argued that greater attention should be paid to the rental housing sector. The statistical link between housing tenure band the level of economic development is less than clear.

Quantitative importance of renting by country and City

Despite the years of effort and financial expenditure that so many governments have spent in trying to expand homeownership, rental housing still constitute a large component of the housing stock in many countries. While the incidence of renting varies considerably across the world, rental housing accommodates a significant share of families in some countries.

Table 1.1: Housing tenure for selected cities, percent (1994-2001)

City	Country	year	Ownership	Renting	Other
Bangkok	Thailand	1998	54	41	5
Pusan	Rep. of Korea	1995	72	28	-
Seoul	Rep. of Korea	1995	70	30	-
Istanbul	Turkey	1994	68	32	-

Source: UNCHS, 2001a

The table shows that tenants and other non-owners outnumber homeowners in some of the Asian major cities. Housing in our country should be in a consistent framework because our resources are limited and wastage of resources is not being expected. So there is a need of

house rent modeling as the factors or criteria’s associated with the sustainable and economically potential housing are considered in the model.

1.3.2 Objectives of national housing policy

There are mainly two objectives of our national housing policy (1993) that focus the need of the present study. Objectives are:

      Develop new strategies from time to time to cope with the emerging housing needs and problems in the country

      Undertake action oriented research in all aspects related to housing and foster minimization of cost and rent

1.3.3 National perspective

Large proportions of urban population in our country are tenant. In our country, housing projects are generally instigated according to the zoning regulation. Due to the projection difficulties of rent structure, most of the housing projects do not assure the cost recovery of the projects as well as the projects tend to be up to snuff. 

The factors or criteria’s associated with the sustainable and economically potential housing are often ignored in the contemporary planning practices like zoning. So there is a need of house rent modeling as the factors or criteria’s associated with the sustainable and economically potential housing are considered in the model.

1.4 Research questions

1. What are the general factors (dependent as well as independent) that influence the house rent?
2. What are the specific independent factors that influence the house rent?
3. What is the relation between independent factors and house rent?
4. What are the degrees of dependence of house rent in respect to the independent factors?
5. How the relation can be integrated with the space (map).?
6. What would be the spatial representation of the relation?
7. What would be the location of house according to ones desired rent?
8. What would be the house rent for an existing location?

1.5 Literature review

1.5.1 Factor analysis

Factor analysis attempts to identify underlying variables, or factors, that explain the pattern of correlations within a set of observed variables. Factor analysis is often used in data reduction to identify a small number of factors that explain most of the variance observed in a much larger number of manifest variables. The factor analysis procedure offers a high degree of flexibility:  

1.5.1.1 Factor analysis data considerations

The variables should be quantitative at the interval or ratio level (SPSS tutorial). Categorical data (such as religion or country of origin) are not suitable for factor analysis. Data for which Pearson correlation coefficients can sensibly be calculated should be suitable for factor analysis.

 

1.5.1.2 Factor analysis assumptions

The factor analysis model specifies that variables are determined by common factors (the factors estimated by the model) and unique factors (which do not overlap between observed variables); the computed estimates are based on the assumption that all unique factors are uncorrelated to each other and with the common factors.

 

1.5.1.3 Factor extraction methods

There are seven method of extraction of the factors. Followings are the description of those methods in brief.

i. Principal Components Analysis

A factor extraction method used to form uncorrelated linear combinations of the observed variables. The first component has maximum variance. Successive components explain progressively smaller portions of the variance and are all uncorrelated with each other. Principal components analysis is used to obtain the initial factor solution. It can be used when a correlation matrix is singular.

ii. Un weighted Least-Squares Method

A factor extraction method that minimizes the sum of the squared differences between the observed and reproduced correlation matrices ignoring the diagonals.

iii. Generalized Least-Squares Method

A factor extraction method that minimizes the sum of the squared differences between the observed and reproduced correlation matrices. Correlations are weighted by the inverse of their uniqueness, so that variables with high uniqueness are given less weight than those with low uniqueness.

iv. Maximum-Likelihood Method

A factor extraction method that produces parameter estimates that are most likely to have produced the observed correlation matrix if the sample is from a multivariate normal distribution. The correlations are weighted by the inverse of the uniqueness of the variables, and an iterative algorithm is employed.

v. Principal Axis Factoring

 A method of extracting factors from the original correlation matrix with squared multiple correlation coefficients placed in the diagonal as initial estimates of the communalities. These factor loadings are used to estimate new communalities that replace the old communality estimates in the diagonal. Iterations continue until the changes in the communalities from one iteration to the next satisfy the convergence criterion for extraction.

vi. Alpha

A factor extraction method that considers the variables in the analysis to be a sample from the universe of potential variables. It maximizes the alpha reliability of the factors.

vii. Image Factoring

 A factor extraction method developed by Guttman and based on image theory. The common part of the variable, called the partial image, is defined as its linear regression on remaining variables, rather than a function of hypothetical factors.

 

1.5.1.4 Factor analysis rotation

There are five method of extraction of the factors. Followings are the description of those methods in brief.

i. Varimax Method

An orthogonal rotation method that minimizes the number of variables that have high loadings on each factor. It simplifies the interpretation of the factors.

ii. Direct Oblimin Method

 A method for oblique (nonorthogonal) rotation. When delta equals 0 (the default), solutions are most oblique. As delta becomes more negative, the factors become less oblique.

iii. Quartimax Method

A rotation method that minimizes the number of factors needed to explain each variable. It simplifies the interpretation of the observed variables.

iv. Equamax Method

A rotation method that is a combination of the varimax method, which simplifies the factors, and the quartimax method, which simplifies the variables. The number of variables that load highly on a factor and the number of factors needed to explain a variable are minimized.

v. Promax Rotation

An oblique rotation, which allows factors to be correlated. It can be calculated more quickly than a direct oblimin rotation, so it is useful for large datasets.

1.5.1.5 Estimating factor score coefficients

There are three methods of estimating factor score coefficient. The methods are below:

i. Show detailsRegression Method

A method for estimating factor score coefficients. The scores produced have mean of 0 and a variance equal to the squared multiple correlation between the estimated factor scores and the true factor values. The scores may be correlated even when factors are orthogonal.

ii. Bartlett Scores

A method of estimating factor score coefficients. The scores produced have a mean of 0. The sum of squares of the unique factors over the range of variables is minimized.

iii. Anderson-Rubin Method

A method of estimating factor score coefficients; a modification of the Bartlett method which ensures orthogonality of the estimated factors. The scores produced have a mean of 0, a standard deviation of 1, and are uncorrelated. 

 

1.5.2 Linear Regression

Linear Regression estimates the coefficients of the linear equation, involving one or more independent variables that best predict the value of the dependent variable. Linear regression is used to model the value of a dependent scale variable based on its linear relationship to one or more predictors. The linear regression model assumes that there is a linear, or "straight line," relationship between the dependent variable and each predictor. This relationship is described in the following formula.

The model is linear because increasing the value of the jth predictor by 1 unit increases the value of the dependent by bj units. Note that b0 is the intercept, the model-predicted value of the dependent variable when the value of every predictor is equal to 0. For the purpose of testing hypotheses about the values of model parameters, the linear regression model also assumes the following:

    The error term has a normal distribution with a mean of 0.

    The variance of the error term is constant across cases and independent of the variables in the model. An error term with non-constant variance is said to be heteroscedastic.

    The value of the error term for a given case is independent of the values of the variables in the model and of the values of the error term for other cases.

 

1.5.2.1 Data Considerations for linear Regression

The dependent and independent variables should be quantitative. Categorical variables, such as religion, major field of study, or region of residence, need to be recoded to binary (dummy) variables or other types of contrast variables.

1.5.2.2 Assumptions of linear regression model

For each value of the independent variable, the distribution of the dependent variable must be normal. The variance of the distribution of the dependent variable should be constant for all values of the independent variable. The relationship between the dependent variable and each independent variable should be linear, and all observations should be independent.

1.5.2.3 Variable Selection Methods of linear Regression model

There are several methods for selecting the variables for linear Regression model. Those methods help to specify how independent variables are entered into the analysis. Using different methods, it could be construct a variety of regression models from the same set of variables. Followings are the description of those methods in brief.

i. Enter (Regression)

A procedure for variable selection in which all variables in a block are entered in a single step.

ii. Stepwise

At each step, the independent variable not in the equation which has the smallest probability of F is entered, if that probability is sufficiently small. Variables already in the regression equation are removed if their probability of F becomes sufficiently large. The method terminates when no more variables are eligible for inclusion or removal.

iii. Remove

A procedure for variable selection in which all variables in a block are removed in a single step.

iv. Backward Elimination

A variable selection procedure in which all variables are entered into the equation and then sequentially removed. The variable with the smallest partial correlation with the dependent variable is considered first for removal. If it meets the criterion for elimination, it is removed. After the first variable is removed, the variable remaining in the equation with the smallest partial correlation is considered next. The procedure stops when there are no variables in the equation that satisfy the removal criteria.

v. Forward Selection

A stepwise variable selection procedure in which variables are sequentially entered into the model. The first variable considered for entry into the equation is the one with the largest positive or negative correlation with the dependent variable. This variable is entered into the equation only if it satisfies the criterion for entry. If the first variable is entered, the independent variable not in the equation that has the largest partial correlation is considered next. The procedure stops when there are no variables that meet the entry criterion.

1.5.2.4 Scatter Plots linear regression

Plots are useful for detecting outliers, unusual observations, and influential cases. plot can be made any two of the: dependent variable (DEPENDNT), standardized predicted values (ZPRED), standardized residuals (ZRESID), deleted residuals (DRESID), adjusted predicted values (ADJPRED), Studentized residuals (SRESID), or Studentized deleted residuals (SDRESID). Followings are two types of scatter plot.

i. Produce all partial plots

Scatter plots of residuals of each independent variable and the residuals of the dependent variable when both variables are regressed separately on the rest of the independent variables. At least two independent variables must be in the equation for a partial plot to be produced.

ii. Standardized Residual Plots

Histograms of standardized residuals and normal probability plots comparing the distribution of standardized residuals to a normal distribution.

1.5.2.5 Predicted values for dependent variables

Predicted Values are the values that the regression model predicts for each case. There are several form of predicted values. They are follows:

i. Show detailsUnstandardized

The value the model predicts for the dependent variable.

ii Standardized

A transformation of each predicted value into its standardized form. That is, the mean predicted value is subtracted from the predicted value, and the difference is divided by the

standard deviation of the predicted values. Standardized predicted values have a mean of 0 and a standard deviation of 1.

iii. Adjusted

The predicted value for a case when that case is excluded from the calculation of the regression coefficients.

iv. S.E. of mean predictions

Standard errors of the predicted values are an estimate of the standard deviation of the average value of the dependent variable for cases that have the same values of the independent variables.

1.5.2.6 Distances of regression model

Measures to identify cases with unusual combinations of values for the independent variables and cases that may have a large impact on the regression model. Methods of distance are follows.

i. Mahalanobis

A measure of how much a case's values on the independent variables differ from the average of all cases. A large Mahalanobis distance identifies a case as having extreme values on one or more of the independent variables.

ii. Cook's

A measure of how much the residuals of all cases would change if a particular case were excluded from the calculation of the regression coefficients. A large Cook's D indicates that excluding a case from computation of the regression statistics, changes the coefficients substantially.

iii. Leverage values

Measures the influence of a point on the fit of the regression. The centered leverage ranges from 0 (no influence on the fit) to (N-1)/N.

1.5.2.7 Residuals

The actual value of the dependent variable minus the value predicted by the regression equation. Methods of calculating residuals are follows:

i. Show detailsUnstandardized

The difference between an observed value and the value predicted by the model.

ii. Standardized

The residual divided by an estimate of its standard deviation. Standardized residuals which are also known as Pearson residuals, have a mean of 0 and a standard deviation of 1.

iii. Studentized

The residual divided by an estimate of its standard deviation that varies from case to case, depending on the distance of each case's values on the independent variables from the means of the independent variables.

iv. Deleted

The residual for a case when that case is excluded from the calculation of the regression coefficients. It is the difference between the value of the dependent variable and the adjusted predicted value.

v. Studentized deleted

The deleted residual for a case divided by its standard error. The difference between a Studentized deleted residual and its associated Studentized residual indicates how much difference eliminating a case makes on its own prediction.

1.5.2.8 Linear Regression Statistics

i. R squared change

The change in the R2 statistic that is produced by adding or deleting an independent variable. If the R2 change associated with a variable is large, that means that the variable is a good predictor of the dependent variable.

ii. Part and partial correlations

Displays the zero-order, part, and partial correlations. Values of a correlation coefficient range from –1 to 1. The sign of the coefficient indicates the direction of the relationship, and its absolute value indicates the strength, with larger absolute values indicating stronger relationships.

iii. Collinearity diagnostics

Collinearity (or multicollinearity) is the undesirable situation when one independent variable is a linear function of other independent variables. Eigenvalues of the scaled and uncentered cross-products matrix, condition indices, and variance-decomposition proportions are displayed along with variance inflation factors (VIF) and tolerances for individual variables.

1.5.3 Criteria ranking methods

The simplest method for assessing the importance of weights is to arrange the criteria in orders of importance. Either straight ranking, from most to least important or inverse ranking, from least to most important, can be used. A third method that can be used is the swing weighting technique. This method retrieves a ranking from the decision maker using the end points of the criteria ranges. The result of these methods is an ordinal ranking. There are four methods exist to translate the ordinal ranking into quantitative ranking. Ranking methods briefly described below.

1.5.3.1 Ordinal ranking

An ordinal ranking can be made quantitative by assigning weights to the criteria in different ways, for example by using equal distances between successive criteria, or, by using methods such as: the expected value method, the extreme value method or the random value method. Moreover, multi-criteria methods exist that use the ordinal weights directly to assess a ranking of the alternatives. An example of such a method is the regime method. These methods treat the ordinal ranking as information on the unknown quantitative weights and try to make optimal use of this information.

Assume that a complete ordering of N criteria c1 to cN is given such that c1 is the most important criterion, followed by c2, then c3, etc. until cN, the least important criterion is reached. Let w1, ., wN denote a set of quantitative weights in accordance with this ordering and impose the conditions that the quantitative weights are non-negative and add up to 1. Then the weights are elements of the following set:

The set of weights of a problem with three criteria (c1, c2, c3) that are non-negative and add up to one is represented as triangle ABC in Figure 2.11. If criterion c1 is more important than criterion c2 and criterion c2 is more important than criterion c3, then the set of feasible weights S equals the shaded triangle ADE in Figure 1.1. This shows that the information contained in the ordering is substantial. From the set of weights that are non-negative and add up to 1, only 1/6 proves to remain feasible.

In this section three methods are included for using the information on the set S of feasible weights to produce quantitative weights: the expected value method (Rietveld, 1980, 1984), the extreme value method (Paelinck 1974, 1977) or the random value method (Voogd, 1983).

1.5.3.2 Random weight method

The random value method assumes that the unknown quantitative weights are uniformly distributed within the set of feasible weights S. Therefore the random value method simulates the variation within a group. Different vectors of feasible weights can result in different rankings of the alternatives. The method uses a random generator to determine a large number of feasible quantitative weights. These weights result in, possibly, different rankings of the alternatives. These rankings are converted into a frequency table. Then, a summation procedure (Voogd, 1983) is used to translate the probabilities assigned to the rankings to probabilities that a certain alternative will obtain a certain rank number (Janssen, 1992).

The numerical weights are found by random drawings of weights that fulfill two conditions:

1. The value of the weights must correspond to the ordinal ranking entered by the user.

2. The sum of weights of all effects is one.

The random weight method always leads to complete ranking. The ranking order is not always in agreement with all the entered quantitative weights, and is therefore not entirely certain. Combined with a multicriteria method, this method leads to a ranking order of all the alternatives. There is no intermediate result in the form of one set of quantitative weights.

1.5.3.3 Expected value method

Similar to the random weight method, the expected value method assumes that each set of weights that fits the rank order of criteria has equal probability. The weight vector is calculated as the expected value of the feasible set (Rietveld 1984). Therefore, the expected value method simulates the average idea/opinion of a group. In the shaded triangle ADE (set S) of Figure 2.11 the expected value is found as the centroid of this triangle. This method results in a unique weight vector. In combination with, for example, weighted summation it also results in a complete ranking of the alternatives. The expected value method gives rise to a convex relationship between ordinal and quantitative weights: the difference between two subsequent weights is larger for more important criteria.

The expected value method calculates the weight, wk, for criterion k according to formula (2.5) where n is the number of criteria. The weights fit the rank order of criteria defined by set S, meaning that  w1 ≥ w2 ≥ …….wn ≥ 0.

Table 1.2 shows the weight vectors for various numbers of criteria according to formula

Table 1.2: Expected value of criterion weights

Number of criteria N	Expected value of criterion weight
	E(w1)	E(w2)	E(w3)	E(w4)	E(w5)	E(w6)
2	0.75	0.25
3	0.61	0.28	0.11
4	0.52	0.27	0.15	0.06
5	0.46	0.26	0.16	0.09	0.04
6	0.41	0.26	0.16	0.10	0.06		0.03

The expected value method, combined with a multicriteria method, always leads to complete ranking. The rank order is not always in agreement with all the possible quantitative weights, the weights of set S, and is therefore not entirely certain.

1.5.3.4 Extreme value method

The extreme value method is based on the extreme quantitative weights, which just fulfill the assigned qualitative ranking of the weights. Because there are always different extreme combinations of weights, this method can result in different rankings of alternatives. Therefore, the extreme value method simulates the extremist/fanatics of a group. In this case it is impossible to determine a complete ranking of the alternatives. In the shaded triangle ADE (set S) of Figure 2.11 the three extreme weight combinations are exactly the corner points A, D and E. The weights of these corner points are presented in Table 1.3.

Table 1.3: Weights for the three extreme points

	Corner A	Corner D	Corner E
Criterion 1	1	0.5	0.33
Criterion 2	0	0.5	0.33
Criterion 3	0	0	0.33

The ranking of the alternatives will be determined for each extreme point (Corner A, Corner D and Corner E). Only orderings of alternatives that are found for all of these corners are

included in the final ranking. The final ranking therefore holds for all feasible values of the weights within triangle ADE of Figure 2.11 and no assumption of the distribution needs to be made. In combination with any of the multicriteria methods described in this chapter, this procedure usually results in an incomplete but certain ranking.

1.5.3.5 Rank sum method

Another method to generate numerical weights from a rank order of criteria is the rank sum method. This method calculates the weight, wk, for criterion k according to formula where n is the number of criteria. Again, the weights fit the rank order of criteria defined by set S, meaning that  w1≥ w2≥……. . ≥ wn≥ 0.

                                         

Table 1.4 shows the weight vectors for various numbers of criteria according to formula

Table 1.4: Criterion weights using rank sum method.

Number of criteria N	Expected value of criterion weight
	E(w1)	E(w2)	E(w3)	E(w4)	E(w5)	E(w6)
2	0.66	0.33
3	0.50	0.33	0.17
4	0.40	0.30	0.20	0.10
5	0.33	0.27	0.20	0.13	0.07
6	0.29	0.24	0.19	0.14	0.10		0.05

The rank sum method, combined with a multicriteria method, always leads to complete ranking. The rank order is not always in agreement with all the possible quantitative weights, the weights of set S, and is therefore not entirely certain. Other methods similar to the rank sum method are the rank reciprocal method and the rank exponent method. Formulas for these methods can be found in Malczewski (1999).

1.5.4 Books

In the book “Rental Housing” published by UN-HABITAT illustrate that renting is anything but a partial answer to the housing problems that so many people in so many settlements both in developed and developing countries are facing. Private owners prefer to rent to members of ethnic groups other than their own because, as they report, it is easier to collect rents from those to whom is not close. . As UN-HABITAT has noted: “governments

should review their housing   policies and device appropriate strategies for rental housing which remove biases against non owners”. The demand for rental housing is determined by the location of the housing as well as the physical condition of the housing.

Harold C. Schuch in the book “GIS data conversion handbook” describes, the data in geographical information system is the model of the real world. By GIS a trainee pilot uses flight simulator, it is, in principle, possible for planners and decision-makers to explore a range of possible scenarios and to obtain an idea of the consequences of a course of action before the mistakes have been irrevocably made in the landscape itself.

Jones, Christopher B. (1997), in his book “Geographical Information Systems and Computer Cartography”, tried to define the term Spatial Decision Making as Making Decisions on the phenomena directly related with space or locations integrating all geo-demographic information from a variety of sources in a spatial context.

1.5.5 Journals

Rapter and Maguire (1992), in the paper “Design Models and functionality in GIS” published in Computers & Geosciences journal, defined GIS Data Model as an information structure which allows the user to store specific phenomena as distinct representations and enables to manipulate the phenomena when held in the system as data.

Tempfli, Klaus (1998), in the paper “3D Topographic Mapping for Urban GIS” published in ITC Journal, has reviewed requirements for an urban GIS and analyses topographic objects, the most important class of urban objects. It has explained adopted data model and the derived polygonal data structure. The structure support full 3D geometry and topology and incorporates raster data for photo-texture.

1.5.6 Article

Michael Batty states in the article “Models in planning: technological imperatives and changing roles” that the functionality within GIS for modeling and related kind of urban analysis is new era of planning approach. GIS are issues involving the organization of planning municipality’s data bases in more general terms from routine planning applications and permits to the archiving of map data that might be used for public as well as professional services. In terms of the professional planning process, GIS has many uses at different scales from the regional to the local level urban planning scale.

1.5.7 Dissertation papers

Ahsanul Kabir stated in his dissertation paper “Development of a decision support model for transit zone location choice; a spatial multicriteria evaluation for Klang Valley, Malaysia” that landuse decisions are typically complex since it has unavoidable trade-offs inherent in protecting or developing specific lands and the differential impact on various stakeholder groups. Stakeholders’ values and interests have to be analyzed to determine the landuse pattern that maximizes consensus or minimizes conflict. In this context he suggests spatial multi criteria evaluation as the analytical tool which can serve to inventory, classify, analyze and conveniently arrange the available information concerning choice-possibilities in spatial planning. 

Shamim Mahabubul Haque stated in his dissertation paper “GIS based models for urban growth and renovation policies” that GIS is an automated tool for the accused, storage, analysis and display geo-graphic data and excellent helping material for geo-spatial decision-making, which would improve planning, and management of different service sectors. Refereeing the dissertation analytical tools of GIS allows the spatial explicit function of buildings. He also explores that in many urban models treatment of space is very arbitrary and abstract, simply because of either incapability of such models in handling spatially explicit geo-information or unavailability of spatially explicit data. The integration of GIS in modeling can overcome these problems.

ASM Shafiqur Rahman in his dissertation “Application of GIS Data Models in Spatial Decision Making on Arsenic Contamination” With the adoption of modern technology and developed analyzing tools, decision-making become comparatively easy and efficient. GIS is one of the tools that can help planners in decision-making on different planning issues with its vast digital storage and automation capabilities. In GIS, real geographical variations are converted into discrete objects to store and analyze in a data model. The data model represents a set of guidelines to convert the real world to the digitally and logically represented spatial objects consisting of the attributes and geometry. The models that are GIS oriented are applicable to determine the policy framework and to evaluate the policies with the time frame.

Billah, Md. Masum (2001), conducted a research titled “Prospects and efficiency of GIS Data Models in Planning Application”, where GIS data models have been analyzed critically to explore its prospect and efficiency along with to show how the data models can be used in planning field. He tried to give a guideline to choose appropriate data model in specific planning applications. In the study he explores that the development control activities can be performed through incorporating the spatial data with aspatial data that it can denote the changing pattern of development activities.

1.6 Limitations

For collecting data at first planned residential area and settlement with vegetation land use have been selected. Then 200m x 200m grids are formed over that selected land use. So the resolution cell size here is 200m x 200m. One house from each grid are surveyed assuming that the sample house represent the whole grid. There may be a variety of houses in each grid. So the house selected for the survey may not represent the whole grid though the dominant housing class was subjected to survey.

The model is based on the planned residential area and settlement with vegetation land use. Due to lack of information, other land use such as unplanned residential area, mixed use etc are excluded in the model. So the model may not fit in that type of land use.

Due to lack of manpower, time and information size of grid are selected as 200m x 200m. This is not quite good resolution for building this type of model. So some error may occur.

One of the limitations in this spatial model is its static nature. The model is based on Khulna City. As the economic, physical, social and other conditions are varying from city to city, so the model may not fit for the other cities of Bangladesh like Dhaka, Chittagong etc.

This study mainly deals with the spatial data for the decision model development. But in reality, other non-spatial information needs to be considered.

2.1 Location

Khulna lay on an important location at the lower extreme of the Ganges delta. It is between 20°38' and 23°1' north latitude and 88°54'and 89°58' east longitude.

2.2 Demographic Characteristics

The main demographic characteristics[i] of the city are:

      Moderately rapid population growth (3.8%), mainly due to rural-urban migration. 

      Population composition characterized by a preponderance of children and low share of adults and aged population, resulting in a high dependence ratio. 

      An imbalance in the gender ratio at 100 males per 118 females. 

      A relatively high literacy rate compared to other cities. 

      Gross population density is very high about 18,000 persons per sq. km. 

      A large proportion of people are engaged in informal activities.

      The family size varies from 4.5 to 5.6

Table 2.1: Population and growth rate of Khulna city, 1901-1998

Year	Population (in ‘000)	Growth Rate (Percentage /Year)
1901	10.4	-
1911	18.1	5.5
1921	23.5	2.7
1931	28.0	1.7
1941	34.0	1.9
1951	42.0	2.1
1961	80.2	6.4
1974	437.3	13.0
1984	561.9	3.5
1991	663.3	1.6
1998	847.5	3.5

 Source: United States Agency for International Development, 1999; Environmental Mapping and Workbook for Khulna city

2.3 Migration

Increases in population in the KCC/Study Area in recent years were largely due to the influx of migrants in the city area as well as in the industrial areas of Khalispur, Daulatpur, Shiromoni and in southern part along the shipyard area. Table 2.2 shows lifetime net migration in Khulna district. It is reasonably assumed that at least 95 percent of the migrants

had selected the KCC and the adjoining industrial and commercial belts as the destinations. In 1961, nearly 43 percent of the population with respect to the Khulna Municipality was migrants- it slightly increased to about 44 percent in 1974. This slow but steady trend of in-migration has continued till the present time, registering, as per present survey of the study area, a migrant household of 48.56 per cent. Within KCC area, the proportion may be well over 50 percent.

Table 2.2: Proportion of life time net migration of Khulna Urban Population 1961 to ‘98

Years	Percentage of urban population
1961	42.94
1974	43.68
1998	48.5

Source: KDA, 1980investment and employment survey of Khulna master plan area.

* This includes Khulna city, Satkhira, Bagerhat and Mongla Port.

** This covers KDA area according to household survey of migrants.Table

2.4 Density and Distribution of Population

Table 2.3 shows the change in the population density per sq. km. during 1961 and 1998 for both KCC and KMA areas. As the city’s core development area the KCC has always a high density of population - the density has nearly doubled since the independence of Bangladesh. On the other hand, the KMA having a large area characterized by rural functional activities registered a much lower overall density of population during the same period. Nevertheless, the increase in the density of population followed the same trend as in the KMA/KCC area (Table 2.3).

Table 2.3: Density of Population per sq. km. in KCC and KMA Area, 196 1-98.

Year	KCC	KMA
1961	1740	707
1974	9506	2585
1981	12216	3602
1991	14420	5090
1998	18424	6504

Sources: BBS (1991).

2.5 Population Projection

Table 2.4: Population Projection of KMPA under Three Assumptions, 190 1-2020

Year	Low Growth Rate			Medium Growth Rate			High Growth Rate
	NI	MR	Total	NI	MR	Total	NI	MR	Total
1998	2.0	1.5	3.5	2.0	1.5	3.5	2.0	1.5	3.5
2000	2.0	1.5	3.5	2.0	2.0	4.0	2.0	2.5	4.5
2005	1.8	1.8	3.6	1.8	2.3	4.1	1.8	2.8	4.6
2010	1.7	2.0	3.7	1.7	2.5	4.2	1.7	3.0	4.7
2015	1.6	1.9	3.5	1.6	2.4	4.0	1.6	2.9	4.5
2020	1.5	1.8	3.3	1.5	2.3	3.8	1.5	2.8	4.3

Base Year 1991. NI = Natural Increase, MR =   Migration Rate

On the basis of these assumptions, three-level population projections for KCC/SMA and the study area and its constituent parts have been made (Table 2.4). In the event of regional development at a faster rate, a high growth rate of population may be anticipated, and the population will reach about 1.3 million by the year 2000, 1:6 million by 2005 and will exceed 2 million in subsequent decade to ultimately reach more than 3.0 million by 2020. This study anticipates this rate of population growth in view of the development strategies being undertaken centering around the proposed EPZ area at

2.6 Employment and income

The present employment situation was found to be not very encouraging in Khulna. Low level of investment is perhaps the main reason. Lack of up to data on employment situation is always a problem for analysis and projection. According to sample survey, 30 percent of the total study area populations (10-69) are in the working age, of which 28 percent were in for male and 2 percent for female. These findings suggest that a huge number of working age population is unemployed. Many of these unemployed are engaged in informal economic activities in the city. After late 60s, the economic condition of the city deteriorated. The economy of a city can best be revealed through the income pattern of the city dwellers.

Table 2.5:   Income label of the people of Khulna city

Monthly income label	Percentage distribution
<2500	14.0
2501-3500	14.0
3501-5500	34.0
5501-10000	30.0
10001<	8.0
Total	100.0

Source: KCC, July 1999; CDS of Khulna city in context on national and city perspectives, Bangladesh

Table 2.6: Occupational structure

Employment opportunity	Percentage distribution
service (both public and private sectors)	14.0
Business	12.0
Hawkers and day laborers	2.0
Self-employed	1.3
Housewives	23.0
Pension holders	2.7
Students	36.0
Others	9.0
Total	100.0

Source: United States Agency for International Development, 1999; Environmental Mapping and Workbook for Khulna city

The average household income per month is Taka 5,543. It is equivalent to per capita yearly income of US$ 283. The Master Plan sets the poverty line at Taka 3,500 as the monthly household income, which in its opinion is "just enough to meet expenditure on food and non-food items.” 

2.7 Spatial linkage

Location and linkage of Khulna City with regional town and growth centers make it the most important city in the region. It is an important inland port and provides good water and rail links to Mongla Port. The development of the city is primarily based on agricultural commodities and non-farm resource base available in the vast hinterland west of the river Padma.

2.8 Commerce and Industries

Since in the ancient time, Khulna was known as a river port cum fishing center. The Shipyard, Newsprint mills, Jute mills and various kinds of factories are situated at the riverside. So, naturally the traditional linear development in riverside is created and Khulna became the 8th largest industrial belt in the world and largest center of trade and commerce in Asia. Khulna is presently the third largest industrial city of Bangladesh. Khulna’s reputation in two industries— salt and molasses were century old.

2.9 The land Resource

Land is scarce resource in Bangladesh, particularly in cities. In Khulna, although, apparently it seems that there is sufficient land for urban development, the reality is that the market is dominated by the private sector, while a large amount of land that is owned by the public sector remained unutilized. The development agencies are unable to capture unearned increment generated due to public intervention in land. Under such circumstances, the land market in the city is not supporting the housing and other urban development.

2.10 Housing

The plan estimated to accommodate 4,64,820 persons in the future Khulna city based on two categories of densities, 70 persons/acre and 25 persons/acre. Eight thousand people more were expected to be accommodated through infilling and flood protection bund. Besides, residential zones were also proposed to impose control over unhealthy residential development. Housing and Settlements Directorate and Khulna Development Authority together executed about 12 site and services housing projects. Besides zoning regulations are found to be fairly adhered. But imposition of development control regulations has been found highly irregular.

2.11 Residential Structure of the City

Central Khulna is the highest density area of the city (i.e. the Ward Number 24, 27). Residential population density varies from 346 to above 480 persons per hectare in the area. However, the overall density of population in Khulna is not that high compared with other metropolitan cities of Bangladesh. The central city residential areas are characterized by private spontaneous development over a longer period of time.

The high-class residential areas were developed mainly through the public/development authority interventions. For instance, parts of Khalishpur, South Daulatpur and middle and east Goalpara, South Central Road, Ahsan Ahnied Road, Mohsin Road, etc. are considered to be old upper class residential areas. New upper class areas are Sonadanga and Nirala developed during the last one and a half decades.

Table 2.7: Existing land use pattern of Khulna City

Land use	Area in Hector	Percentages
Commercial	207.61	1.57
Residential	8237.95	62.29
Health, Education and community	905.21	2.77
Transport and communication	2842.99	21.5
Government	1214.15	3.72
Industry	593.10	4.48
Recreation	42.70	.32
Other	539.65	4.08
Total	13226.05	100

Source: KDA Master Plan Interim Report, 2000

2.12 Transportation System

Transport network is one of the, major factors in determining land use. In Khulna, all three important types of transport system, i.e. road, rail and river, run parallel from south to north maintaining a very close proximity (Fig. 2.2). It is a strategic location for the river port as the Bhairab River has a natural bent here and turns into Rupsha River flowing into the Bay-of-Bengal. So, the river current is gentle allowing the location of’ a port. It also connects the seaport of Mongia through the inland water transport. The river route between Khulna and Mongla plays an important role for regional transportation as welt as local consumption. In order to facilitate this river port the railway existed in this zone since the British period. Therefore, the historic and geographical importance of the railway station as well as the port can not be ignored. Khulna is the centre of the southern region of Bangladesh; therefore, the major inter-city road traffic passes close to this port area, like Khulna-Jessore Road, Satkhira Road, Batiaghata Road, and Khulna Mongla, Road. At present KDA has proposed a by-pass road for the diversion of intercity traffic outside the city. The need for a cargo terminal, which will act as transport interchange, can not be overlooked. Considering the present geographical condition and the functional need• of the region, JICA study has already proposed to establish a Multi-Modal Cargo Terminal at Khulna river port area.

2.13 Drainage Pattern

The drainage pattern surrounding the city puts a serious condition for its growth. The city has limitations for its expansion towards east and west. Beel Dakatia, lower Rupsha and its tributaries along the  Moyur pose physical barriers in the expansion towards the west while in east Bhairab-Rupsha and its tributaries are handicap.both directions land gently slopes down.

Table: 2.9: Existing drains in KCC and the extended area

Area	Length ( km)	Type of drain
		Pucca (km)	Semi pucca (km)	Katcha (km)
KCC	528.12	64.08	51.79	380.75
Extended area	0	0	0	0
Total	528.12	64.08	51.79	380.75

Source: KDA master plan. 1999.

2.14 Recreation

2.14.1 Cinema Hall

There are thirteen cinema halls in K.C.C area. Among them four in C.B.D., two nearer C.B.D., one beyond the K.D.A newmarket, one in Boyra, three in Khalishpur, one in Daulatpur and one in Fulbari gate. Only one of them has air conditioning system.

2.14.2 Auditorium

There are 6 auditoriums in Khulna City. Regular functions take place in Zia hall and Umesh Chandra public library auditorium. Besides in Zilla school auditorium and in Shahid Minar of Hadis Park, drama and cultural activities are performed. B.N.A auditorium and C& B auditorium, are also venues of cultural programmes.

2.14.3 Park

There are nine parks and open spaces in the khulna City. Name of those parks and open spaces are given to the following table

Table 2.10: name of park and open space and their estimated area

Name of parks/open space.	Estimated area (acres)
Hadis park	5.59
Sonadanga R/A park	3.77
Nirala R/A park	1.96
Khalishpur wonderland park.	8.37
Boira R/A park	1.75
Jatisango park.	0.8.
Golokmoni park.	0.46
Prem kanon	1.00
Total	23.70

Source: K. D. A. 1999.

  

Map 2.1: Location of the study area (profile 1)

 

Map 2.2: location of the study area (profile 2)

2.15 Physical Characteristics of Khulna City

Khulna is a unique city. Its uniqueness is reflected in its shape, activities and in many other physical and social characteristics. The following are the notable characteristics of Khulna:

      Linear shape extending from southeast to northwest.

      Fish processing activities in and around the city;

      Mixed land use pattern is dominant.

      Low rise buildings;

      Low net residential density, although the gross population density in KCC area is quite high.

      Flat fertile land;

      Least affected city by natural disasters such as flood, cyclone and earthquake.

      Impact of high salinity;

      Known as Industrial City.

      Urban village character.