evaluasi sistem perangkutan sampah kota jember

TRIP GENERATION ANALYSIS USING MULTIPLE LINEAR REGRESSION METHOD ON BUMI ESTATE MUKTISARI AND TAMAN GADING HOUSING JEMBER REGENCY Sonya Sulistyono Lecture of Civil Engineering Departement, Engineering Faculty, Jember University Jl. Slamet Riyadi 62 Jember [email protected]

Akhmad Hasanuddin Lecture of Civil Engineering Departement, Engineering Faculty, Jember University Jl. Slamet Riyadi 62 Jember [email protected]

Yurike Ogi Adrisyanti Student of Civil Engineering Departement, Engineering Faculty, Jember University Jl. Slamet Riyadi 62 Jember

ABSTRACT By increasing the population in Jember, also increase the need for housing. BWK IV (Tegal Besar Region) is one of the rapidest growing residential areas. There are 18 new housings in this region. So, trip generation studies of some housing in BWK IV need to be done to estimate the number of trips. Bumi Estate Muktisari and Taman Gading Housing is one of highest number of housing, and have a dependency on J1. Letjen Suprapto as the main access road to the central part of Jember. The number of occupancy in two housing reached 2,355 units (Bumi Estate Mutisari is 1,237 units and Taman Gading is 1,228 units). Trip generation analysis based on household that apply the step by step type-2 on those housing. The best trip generation model is Y = - 0.439 + 0.174X2 + 0.741X3 + 0.797X4 + 0.484X6. Independent variables that affect the production of trips include: the number of family members (X2), the family member who worked (X3), the family member who get school (X4), and vehicle ownership (X6). The R2 value is 0.74 which means that 74% of production is influenced by the free variables on that model. Keywords: multiple linear regression, trip generation, home-based trip

INTRODUCTION Jember area is 3,293.34 km2 or the third largest area after Banyuwangi and Malang regency. Based on the results of the census on 2010, the population growth in Jember is second highest after Surabaya. It is 2,329,929 inhabitants. The population density average is 707 inhabitants/km2 (BPS, 2010). Jember, one of the areas in the eastern province of East Java, has evolved into a center of various activities such as a government service centers, an education development center and a regional economic service centers (districts) and local (city). This condition causes Jember developed rapidly. By the increasing of population growth, the need for housing also increased. The high price of land in urban centers led to residential land has increasingly shifted to the suburban. While the location of employment and education tend to be more concentrated in urban centers. This is one of the reasons that led to the expansion of the city. One form of the expansion of the city in Jember is the emergence of new residential housing in BWK IV (Tegal Besar). There are 18 housings with multiple developers. The emergence of these estates will surely cause trip generation that will affect the performance and intersections of roads. Related research ever done is a review of the trip attraction (Soetjipto and Sulistyono, 2006). This study reviewed the trip attraction that occurred in

1

The 15th FSTPT International Symposium, STTD Bekasi, November 23-24, 2012

the downtown Jember in BWK I. From the results of these studies show the main roads used by society towards BWK I of IV is Jl. Basuki Rahmad and Jl. Letjend Suprapto. Bumi Estate Muktisari and Taman Gading is a residential area in BWK IV that built by the developer with the largest number of housing units. Bumi Estate Muktisari is 1,127 units. On the other side, Taman Gading consists of 1,228 units. The existence of this housing will cause trip generation that affect the capacity and performance of the main road J1. Basuki Rahmat and J1. Letjend Suprapto. Trip generation study is performed to estimate the household-based trip generation that arising from the residential area. The expected results of this study are to be useful in estimating the amount of assistance from BWK IV in real estate development in the future and it can be seen how much effect on the capacity of the road network in the city of Jember.

LITERATURE REVIEW Trip Generation Model According to Tamin (2003:40), transportation model performed in 4 (four) stages that is continuous. It usually called Four Steps Model that consists of trip generation, trip distribution, modal split and trip assignment. Trip generation consists of trip production and trip attraction. Trip production is used for a home-based movement that has its place of origin or destination. It is home or movement that generated by the not-home-based movement. While the trip attraction is used for a home-based movement, that has its place of origin or destination not a house or a movement that pull up by home-based movement. Trip generation usually uses zone-based data to model the amount of movement that occurs (either generation or pull), such as land use, vehicle ownership, population, employment, population density, income, and also the mode of transportation used. Trip generation should be analyzed separately by the pull of the movement. Thus, the ultimate goal of the movement is to estimate the pull of the movement and trip generation accurately in the present which will be used to predict the future movement (Tamin, 2003:112). Multiple Linear Regression Analysis This concept is a further development of the linear regression, especially in the case that ^ has more independent variables and the parameter b . It is really needed in the reality that shows several land use variables stimulant affect the movement generation. The general form of multiple linear regressions analysis equation method is:

Y  A  B1 X 1  B2 X 2  . . .  BZ X Z

................

(1)

Where : Y = dependent variable X1…Xz = independent variables A = regression constanta B1...Bz = regression coefficients Multiple linear regression analysis is a statistical method. To use it, there are several assumptions:

2


a. Variable values, especially independent variables have a specific value or a value derived from the survey results with no significant errors. b. Dependent variable (Y) must have a linear correlation with independent variables (X). If the relationship is not linear, linear transformations must do although this restriction will have other implications in the analysis of residuals. c. Independent variables effect on dependent variable is the sum, do not have to have a strong correlation among independent variables. d. Variations of dependent variable towards the regression line same for all values of independent variables. e. Variations of dependent variable have to scattered normally or near-normal minimum. f. Value of independent variables is a scale that is relatively easy to be projected. Some things that should be considered, including: a. Multikolinear, this happens because of the linear relationship between variables. Some equations containing the regression coefficients ( ) are not independent each other and cannot be uniquely solved. b. Determination coefficient, the same shape with equation (2). Value of b additional variables typically increase the value of R, to overcome the R value has been corrected.  ( N  1)   ( N  K  1)   

K   R  R 2  N  1   2

................

(2)

c. Corelation coefficient, used to determine the correlation between dependent variable and independent variables or among independent variables. Correlation coefficient can be calculated in various ways, one of them is the equation (3) below: N

r

N

N

i 1

i 1

N  ( X i Yi )   ( X i )   (Yi ) i 1

 N  N  N  ( X i )2    ( X1 )   i 1   i 1

2

2   N  N   2    N  (Y1 )    (Yi )    i 1     i 1

................

(3)

Value of r = 1 means that the correlation between the variables Y and X is positive. Conversely, if the value of r = -1, the correlation between variables Y and X is negative (the increase of X value will affect the decrease of Y value). Value of r = 0, indicates no correlation between the variables. d. T-test, used to test the significance of the correlation coefficient (r) and test the significance of regression coefficients. Any variables that have a regression coefficient is not statistically significant should be removed from the model.

RESEARCH METHODOLOGY The purpose of this research is to get the best model that can estimate the number of trip generation on Bumi Estate Muktisari and Taman Gading housing using the multiple linear regression method by the approach analysis of step-by-step type 2. The primary data obtained from questionnaires distributed to the study sites, and secondary data obtained

3


from the office of the Public Works Human Settlements Jember, Statistics Jember and housing developers of Bumi Estate Muktisari and Taman Gading. Secondary data include: population, residential housing in Jember, residential site plan study area. Preliminary test conducted on 15 respondents from Bumi Estate Muktisari and 15 respondents from Taman Gading. The result of the validity test obtained correlation numbers between 0.44 to 0.97 and has met with significant numbers r Table 0.361. Test reliability with Alpha Cronbach's method, an instrument said to be reliable if an alpha value greater than 0.60 (Ghozali, 2004:42). Calculations on preliminary test results obtained alpha croanbach value (α) of 0.973 > 0.60. It can be concluded that these statements qualify respondents reliability. Production data traveling on the preliminary test is then used as the basis for calculating the sample. With a 95% confidence level specifications, the possibility of sampling error not more than 5% from the sample mean. Based on the statistical tables, there is 1.96 points from the standard error to level confidence (z) 95%. In order to get an acceptable error of no more than 5%, so the number of data samples should be calculated as follows: Sampling error (Se) is acceptable

Se ( x)



Se z



0,195 1,96



= 0,05 x average trip production = 0,05 x 3,9 trip/family/day = 0,195

0,09946

The large number of samples can be calculated: s2 n' n'  for unlimited data and n  for limited data 2 Se( x) 1  n' N Then: n’ = (1,26899)2 / [0,09949]2 = 162,69 ≈ 163 (for unlimited data) n = 163 / (1 + 163/2355) = 152,4 ≈ 152 (for unlimited data) Based on the preliminary test and the population in the study area obtained the minimum sample is 152 samples. In the implementation of this study, respondents were interviewed in the home interview survey was of 203 respondents. As many as, 44% respondents of the Bumi Estate Muktisari and 56% of respondents from Taman Gading.

RESULTS AND DISCUSSION Description of Result Survey Home interview survey conducted in 203 households, 44% of respondents from the Bumi Estate Muktisari and 56% of respondents from Taman Gading. Age head of household between 41 - 50 most dominating. All respondents of 38.9% by the end of graduate school education (48.3%). Most respondents have any kind of civil work/ state (39.9%) with 29.1% monthly income by 1 - 3 million and ownership of the vehicle 2 units (34.5%). A total of 46.3% of respondents traveling for the purpose of education. Preliminary estimates, the most dominant variable affecting the production of the trip was the family member who attended.

4


Statistical Analysis Validity and Reliability of Data Test The results of testing the validity can be seen in Table 1. For a significance level of 5% points of criticism is 0.13775. Value of r can be seen in the table, the number of correlations obtained from variable X1 to variable X6 to Y is just above the 5% level of criticism. Then the respondents had significant statement. Tabel 1 Validity results test Corelation

Variable

r tabel

Note

0.643

0.13775

Valid

0.437

0.884

0.13775

Valid

Number of family members (X2)

0.652

0.453

0.13775

Valid

Family members working (X3)

0.449

0.430

0.13775

Valid

Family members attending school (X4)

0.660

0.422

0.13775

Valid

Income (X5)

0.556

0.865

0.13775

Valid

Vehicle ownership (X6)

0.728

0.765

0.13775

Valid

Y

TOTAL

Trip production (Y)

1.000

House type (X1)

For the ity test, based on correlations between variables obtained alpha croanbach calculation as follows: K *r 7 x 0,465917 9,784   0,859285 , with r   0,465917 1  ( K  1) r 1  (7  1) 465917 21 It can be concluded that all the variables in this study is reliable because it has an alpha croanbach value (α) > 0.60.



Outlier dan Normality Test

Figure 1 Normal Distribution Graph

5


Based on outlier test results, then there are 14 data obtained from 203 data outliers or deviant. These data are not used in further analysis. Based on Figure 1, illustrates that the data is located and scattered around or approaching along the diagonal line. This indicates that the data was normal distribution. Multiple Linear Regression Test Multiple linear regression analysis in this study is using the Step By Step method Type 2 (Tamin 2003: 126). These methods reduce the number of independent variables gradually to obtain the best model which only consists of several independent variables. The results of the regression test iterations are shown in Table 2. Autocorelation and Multicolinierity Test To review the occurs of autocorrelation, is done by the Durbin Watson test. If the values are in the range of - 2 < Durbin Watson < 2, then there is no autocorrelation. Test results obtained Durbin Watson value is 1.596. It can be concluded that there is no autocorrelation among independent variables (- 2 < 1.596 < 2). Table 2 Iteration result of regression test No

Model

R2

Regression Coefficient Constanta

X1

X2

X3

X4

X5

X6

1

Model-1

-0,419

-0,002

0,182

0,732

0,803 0,007 0,426

0,739

2

Model-2

-0,463

-

0,182

0,730

0,802 0,006 0,418

0,740

3

Model-3

-0,439

-

0,174

0,741

0,797

-

0,484

0,740

4

Model-4

-0,082

-

-

0,817

0,911

-

0,512

0,735

5

Model-5

0,125

-

-

1,135

1,197

-

-

0,688

6

Model-6

1,770

-

-

-

1,209

-

-

0,504

From Table 3 above, obtained VIF value of each independent variable < 10 and tolerance values > 0.1. It can be concluded that there is no multicollinearity. Tabel 3 Test results of multicolinearity test Model (Constant)

Tolerance

VIF

X1

0,661

1,513

X2

0,436

2,294

X3

0,685

1,460

X4

0,418

2,390

X5

0,518

1,930

X6

0,335

2,983

From Table 4, the magnitude of correlation between the independent variables appears that the only variable X4 (a member who get school) that have the greatest number of correlations between number other correlation that is equal to - 0.570 or about 57%.

6


Because of this correlation is still below 95%, it can be said that there is a serious multicollinearity (Sugiyono, 2002:93). Tabel 4 Correlation coefficient Variable Correlations

X6

X3

X1

X2

X5

X4

X6

1,000

- 0,266

- 0,213

- 0,170

- 0,475

- 0,339

X3

- 0,266

1,000

- 0,054

- 0,320

- 0,066

0,397

X1

- 0,213

- 0,054

1,000

0,008

- 0,257

- 0,046

X2

- 0,170

- 0,320

0,008

1,000

0,073

- 0,570

X5

- 0,475

- 0,066

- 0,257

0,073

1,000

0,054

X4

- 0,339

0,397

- 0,046

- 0,570

0,054

1,000

Heteroscedasticity Test Can be seen in Figure 2, the points spread randomly above and below the 0 on the y-axis this indicates that there is no heteroscedasticity or otherwise occurred homoskesdastisitas.

Figure 2 Scatterplots Graphs F-test dan T-test According to the test results show that Fcount > Ftable is 89.64 > 2.42 and a significance smaller than the 0.000 < 0.005. It can be concluded that H0 is rejected and HA is accepted. House type variable, number of family members, family member who worked, schooling family members, income, and vehicle ownership influence together significantly and have a linear relationship on the trip production. Meanwhile, according to the t-test results in Table 5, shows that the independent variable type (X1) and income (X5) had no significant effect on the production of trip (Y), as shown by the tcount < ttable and significance rate > 0.05.

7


Tabel 5 T-test result Independent Variable

Tcount

ttable

Sig.

X1

-0.322

1.94318

0.748

X2

2.099

1.94318

0.037

X3

6.158

1.94318

0.000

X4

8.207

1.94318

0.000

X5

1.157

1.94318

0.249

X6

3.911

1.94318

0.000

The Best Model Based on the test results of multiple linear regression analysis of the two methods can be chosen the best linear regression model equation as follows: Y  0,439  0,174 X 2  0,741X 3  0,797 X 4  0,484 X 6  e . . . . . . . . . . . . . .

(4)

Interpretation of the regression equation (Equation 4) above can be expressed as follows: a. Constants of - .439 indicates the magnitude of production at the way the variable type, number of family members, family members and school work, income, and vehicle ownership equals zero. So in this case the value of the independent variable is the amount of traveling production of -0.439. b. Value of b2 = 0.174, meaning that the variable X2 (number of family members) will affect the trip production assuming the variable X3, X4, and X6 are constant then any increase of the number of family members variable (X2) of 1 would increase production by 0.174 trip. c. Value of b3 = 0.741, means that the variable X3 (family member who worked) will affect the trip production, with assuming the variable X2, X4, and X6 constant then any increase family member working variable (X3) of 1 would increase production by 0.741 trip. d. Value of b4 = 0.797, means that the variable X4 (a member who get school) will affect the traveling production assuming the variable X2, X3, and X6 constant then any increase family member schools variable (X4) of 1 would increase production by 0.797 trip. e. Value of b6 = 0.484, means that the variable X6 (vehicle ownership) will affect the traveling production assuming the variable X2, X3, and X4 constant then any increase in vehicle ownership variables (X6) of 1 would increase production by 0.484 trip. Determination Coefficient (R2) The coefficient of determination is a measure of how far the core ability of the model in explaining the variation in the independent variable. From the results of the regression test, obtained the coefficient of determination (R2) of 0.740. This means that the independent variables (number of family members, members of working families, members of the school family and vehicle ownership) is quite influential in the model by 74% to the

8


dependent variable (production assistance) while the remaining 26% is due to other factors not included in the regression model.

CONCLUSION From the analysis and discussion are carried out in determining the trip generation models and the number of trips on Bumi Estate Muktisari and Taman Gading can be concluded that the selection of the best model using the step-by-step type 2 obtained the best model is Y = - 0,439 + 0,174X2 + 0,741X3 + 0,797X4 + 0,484X6. Which means 74% of production is influenced by the way the independent variables in the model include: the number of family members (X2), a family member who worked (X3), a member who get school (X4), and vehicle ownership (X6).

REFERENCES Agusyana, Yus. 2011. Olah Data Skripsi dan Penelitian dengan SPSS 19. Jakarta : PT. Elex Media Komputindo. Anonim. 2010. Hasil Sensus Penduduk 2010 Data Agregat Per Kecamatan Kabupaten Jember. Jember : Badan Pusat Statistik Kabupaten Jember. Ghozali, Imam. 2004. Aplikasi Analisis Multivarial dengan Program SPSS. Semarang : Badan Penerbit Universitas Diponegoro. http://repository.usu.ac.id/bitstream/123456789/26009/1/Appendix.pdf, “Kuisioner Data Analisa Bangkitan Perjalanan Pada Kecamatan Deli Tua”. diakses tanggal 28 November 2011. www.google.com. Lubis, M.E. 2008. “Penetapan Model Bangkitan Pergerakan Untuk Beberapa Type Perumahan Di Kota Pematangsiatar (Studi Kasus: Perumahan Pinggiran Kota Pematangsiatar)”. Tesis. Sumatera Utara : Universitas Sumatera Utara. Sekaran, Uma. 2006. Metodologi Penelitian untuk Bisnis Edisi 4 Buku 2. Jakarta : Salemba Empat. Singarimbun, Masri and Sofian. 1995. Metode Penelitian Survai. Cetakan kedua. Jakarta : LP3S. Soetadji, R.D., 2008, “Pemodelan Bangkitan Pergerakan Untuk Menentukan Jumlah Perjalanan Dengan Metode Analisis Regresi Di Kota Jember”. Skripsi. Jember : Universitas Jember. Soetjipto, Jojok Widodo. and Sulistiyono, Sulistyono, 2006, “Model Atraksi Perjalanan (Trip Attraction Model) Pada Wilayah Pusat Kota Jember (Eks.Kotip Jember)”, Jurnal Rekayasa Edisi 2 Tahun 3, Jember : Fakultas Teknik Universitas Jember. Sugiyono. 2002. Metode Penelitian Bisnis. Bandung : Alfabeta. Tamin, O.Z. 2003. Perencanaan dan Permodelan Transportasi. Edisi Kedua. Bandung : Penerbit ITB.

9


Trihendradi, C. 2009. Step by Step SPSS 16 Analisis Data Statistik. Yogyakarta : Penerbit ANDI.

10