Published: 28 July 2016 Author(s) 2016. CC-BY 3.0 License ... · 2 Faculty of Engineering, Universitas Diponegoro, Jalan Prof. H. Soedarto, SH, Tembalang, Kota Semarang - 50275, Indonesia

Hydrochemical assessment of Semarang area usingmultivariate statistics: A sample based datasetIrawan Dasapta Erwin1 and Putranto Thomas Triadi2

1Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Jalan Ganesa No. 10,Bandung - 40132, Indonesia2Faculty of Engineering, Universitas Diponegoro, Jalan Prof. H. Soedarto, SH, Tembalang, KotaSemarang - 50275, Indonesia

Correspondence to: Dasapta Erwin Irawan ([email protected])

Abstract. The following paper describes in brief the data set related to our project "Hydrochemical

assessment of Semarang Groundwater Quality". All of 58 samples were taken in 1992, 1993, 2003,

2006, and 2007 using well point data from several reports from Ministry of Energy and Mineral

Resources and independent consultants. We provided 20 parameters in each samples (sample id,

coord X, coord Y, well depth, water level, water elevation, TDS, pH, EC, K, Ca, Na, Mg, Cl, SO4,5

HCO3, year, ion balance, screen location, and chemical facies). The chemical composition were

tested in the Water Quality Laboratory, Universitas Diponegoro using mas spectrofotometer method.

The statistical treatment for the dataset (available on Zenodo doi:10.5281/zenodo.57293) were

described as follows: (1) data preparation in to csv file format, load it in to R environment; (2)

data treatment, including: correlation matrix, cluster analysis using kmeans and hierarchical cluster10

analysis, and principal component analysis. For analysis and visualizations, We used the following

R packages: ggplot2, dplyr, factomineR, factoExtra, cluster, ggcorrplot, and

ape.

1 Introduction

The following paper describes in brief the data set related to our project "Hydrochemical assessment15

of Semarang Groundwater Quality". The aim of this project is to understand the water quality clas-

sification and distribution in Semarang area and to explain the underlying processes. This analysis

is very important with the vast development of infrastructure (Putranto and Rüde (2016)) and urban

settlement in coastal area and the rate of salinity encroachment (Rahmawati and Marfai (2013)). The

location of the study is Semarang area, Indonesia.20

1

Earth Syst. Sci. Data Discuss., doi:10.5194/essd-2016-29, 2016

Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns

Manuscript under review for journal Earth Syst. Sci. DataPublished: 28 July 2016c© Author(s) 2016. CC-BY 3.0 License.

2 General description of the dataset

2.1 Samples

All of 58 taken in 1992, 1993, 2003, 2006, and 2007 in 1992, 1993, 2003, 2006, and 2007 using well

point data from several reports from Ministry of Energy and Mineral Resources and independent

consultant. We provided 20 parameters in each samples: sample id, coord X, coord Y,25

well depth, water level, water elevation, TDS, pH, EC, K, Ca, Na, Mg,

Cl, SO4, HCO3, year, ion balance, screen location, and chemical facies.

The chemical composition were tested in the Water Quality Laboratory, Universitas Diponegoro us-

ing mass spectrofotometer method. The laboratory procedures followed the SNI (Indonesia National

Standard) for water quality measurement (BSN (2012)), which is comply to the US-EPA standards.30

The original dataset is available on Zenodo (Irawan and Putranto (2016)).

Figure 1. The location of well point and the Stiff diagram

2.2 The value of dataset

The following list describes the value of the dataset:

– It provides the current setting of water quality as the baseline of environmental monitoring of

the area and serves as a source of groundwater quality indicator for the regional planning of35

the area,

– It promotes the importance of open government dataset and enriches the library of water qual-

ity dataset of the area,

– It sets an example of data re-use and re-analysis in hydrogeological research landscape.

2


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


3 Geographical coverage40

The sampling area is Semarang area, the capital of Mid Java Province, Java, Indonesia. The sampling

points were distributed from the southern volcanic highland to the coastal area. The coordinate of the

area is (420000, 9240000) and (470000, 9220000). We plotted the data points using UTM-WGS84-

48S projection system.

4 Statistical design45

The hierarchical cluster analysis (HCA) and principal component analysis (PCA) are both widely

used in the hydrochemical analysis (Adams et al. (2001); King et al. (2014); Ayenew et al. (2009);

Deon et al. (2015); Wilkinson (2014); Maechler et al. (2016)). We have applied the two approaches

on groundwater in volcanic area on various locations (Irawan et al. (2009); Herdianita et al. (2010)).

The R implementation was based on Coghlan (2009).50

4.1 Data preparation

The dataset was formatted in the csv (comma separated value) before parsed in to R program (R

Core Team (2016)) for analysis using the following R packages: ggplot2 (Wickham (2009)), dplyr

(Wickham and Francois (2016)), factomineR (Lê et al. (2008)), factoExtra (Kassambara and Mundt

(2016)), cluster (Maechler et al. (2016)), ggcorrplot (Kassambara (2016)), and ape (Paradis et al.55

(2004)).

d f <− as . d a t a . f rame ( r e a d . csv ( " data_smg . csv " ) ) # l o a d i n g as d a t a f rame

head ( d f ) # c h e c k i n g h e a d e r

i s . na ( d f ) # c h e c k i n g NAs i n d f

df2 <− df [ c ( 2 , 5 : 1 8 ) ] # s u b s e t t i n g df , e x c l u d e v a r wi th NAs60

head ( df2 )

i s . na ( d f2 ) # c h e c k i n g NAs i n df2

s t r ( d f2 ) # c h e c k i n g d a t a t y p e i n df2

i s . numer ic ( d f2 ) # c h e c k i n g d a t a t y p e i n df2

rownames ( df2 ) <− d f 2 $ l o c a t i o n # s e t t i n g c o l l o c a t i o n as row names65


4.2 Data treatment

The dataset was treated using the following method: correlation matrix, HCA, and PCA. the steps

and R code can be described below.

3


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


4.2.1 Correlation matrix70

Here we used PerformanceAnalytics and ggcorrplot packages to build a correlation ma-

trix. The following is the code.

## u s i n g P e r f o r m a n c e A n a l y t i c s

i n s t a l l . p a c k a g e s ( " P e r f o r m a n c e A n a l y t i c s " )

l i b r a r y ( P e r f o r m a n c e A n a l y t i c s )75

c h a r t . C o r r e l a t i o n ( df2 , h i s t o g r a m =TRUE, pch =19) # v i s u a l PA

## u s i n g g g c o r r p l o t

i n s t a l l . p a c k a g e s ( " g g c o r r p l o t " )

l i b r a r y ( g g c o r r p l o t )80

c o r r e l <− round ( c o r ( d f2 ) , 1 ) # r o u n d i n g c o r r e l m a t r i x

head ( c o r r e l [ , 1 : 1 4 ] ) # view h e a d e r s

p . mat <− cor_pmat ( d f2 ) # compute p−v a l u e s

head ( p . mat [ , 1 : 1 4 ] ) # view h e a d e r s

g g c o r r p l o t ( c o r r e l ) # making heatmap85

4.2.2 Hierarchical cluster analysis (CA)

We build the CA using k-means and hierarchical clustering by implementing R base function and

factoextra package, based on the following code.

i n s t a l l . p a c k a g e s ( " f a c t o e x t r a " )

# i n s t a l l _ g i t h u b ( " kas sambara / f a c t o e x t r a " )90

i n s t a l l . p a c k a g e s ( " c l u s t e r " )

l i b r a r y ( c l u s t e r )

l i b r a r y ( f a c t o e x t r a )

### k means method95

km2 <− kmeans ( df2 , 2 , n s t a r t = 25) # kmeans wi th 2 c e n t e r s

km3 <− kmeans ( df2 , 3 , n s t a r t = 25) # kmeans wi th 3 c e n t e r s

k m 2 $ c l u s t e r # e x t r a c t i n g c l u s t e r number

km2$cen t e r s # e x t r a c t i n g c l u s t e r means ( o r c e n t e r s )

p lo tkm2 <− p l o t ( df2 ,100

c o l = k m 2 $ c l u s t e r ,

pch = 19 ,

f rame = T ,

main = "K−means wi th k = 2 " ) # n o t e s : need l o n g e r x a x i s

4


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


p o i n t s ( km2$cen te r s ,105

c o l = 1 : 2 ,

pch = 8 , cex = 3)

k m 3 $ c l u s t e r # e x t r a c t i n g c l u s t e r number

km3$cen t e r s # e x t r a c t i n g c l u s t e r means ( o r c e n t e r s )110

plotkm3 <− p l o t ( df2 ,

c o l = k m 3 $ c l u s t e r ,

pch = 19 ,

f rame = T ,

main = "K−means wi th k = 3 " )115

p o i n t s ( km3$cen te r s ,

c o l = 1 : 2 ,

pch = 8 ,

cex = 3)

120

### e v a l u a t i n g c l u s t e r

d f2 <− s c a l e ( d f2 )

head ( df2 )

f v i z _ n b c l u s t ( df2 ,

kmeans , method = " wss " ) +125

geom_vl ine ( x i n t e r c e p t = 3 ,

l i n e t y p e = 2) # d e t e r m i n i n g o p t i m a l no c l u s t e r

km3 . r e s <− kmeans ( df2 , 3 , n s t a r t = 25) # r u n n i n g kmeans wi th 4 c l u s t e r

p r i n t ( km3 . r e s ) # p r i n t o u t p u t

f v i z _ c l u s t e r ( km3 . r e s , d a t a = df2 ) # v i s o u t p u t130

pam . r e s <− pam ( s c a l e ( d f2 ) , 3 ) # r u n n i n g pam c l u s t e r w i th 3 c l u s t e r

pam . r e s $ m e d o i d s # e x t r a c t medoids

c l u s p l o t ( pam . r e s ,

main = " C l u s t e r p l o t , k = 3 " ,135

c o l o r = TRUE)

p l o t ( s i l h o u e t t e ( pam . r e s ) , c o l = 2 : 5 )

f v i z _ s i l h o u e t t e ( s i l h o u e t t e ( pam . r e s ) )

c l a r a x <− c l a r a ( df2 , 3 , s ample s = 5) # u s i n g c l a r a method

f v i z _ c l u s t e r ( c l a r a x ,140

s t a n d = FALSE ,

5


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


geom = " p o i n t " ,

l a b e l =T ,

p o i n t s i z e = 1)

145

### C r e a t i n g dendogram

d i s t d f 2 . r e s <− d i s t ( df2 ,

method = " e u c l i d e a n " )

h ca d f 2 <− h c l u s t ( d i s t d f 2 . r e s ,

method = " c o m p l e t e " )150

p l o t ( hcadf2 ,

hang = −1) # dendogram v i s

r e c t . h c l u s t ( hcadf2 ,

k = 3 ,

b o r d e r = 2 : 4 ) # dendogram v i s wi th g r o u p i n g155

### u s i n g n b c l u s t pack t o e v a l u a t e no of c l u s t e r

i n s t a l l . p a c k a g e s ( " NbClus t " ) # f o r more p r e c i s e no of c l u s t e r

l i b r a r y ( " NbClus t " )

r e s d f 2 . nb <− NbClus t ( df2 ,160

d i s t a n c e = " e u c l i d e a n " ,

min . nc = 2 , max . nc = 10 ,

method = " c o m p l e t e " ,

i n d e x =" gap " )

r e s d f 2 . nb # p r i n t t h e r e s u l t s165

r e s d f 2 . nb$Al l . i n d e x # A l l gap s t a t i s t i c v a l u e s

r e s d f 2 . nb$Bes t . nc # B es t number o f c l u s t e r s

r e s d f 2 . nb$Bes t . p a r t i t i o n # c a l c u l a t e b e s t p a r t i t i o n

nbdf2 <− NbClus t ( df2 ,

d i s t a n c e = " e u c l i d e a n " ,170

min . nc = 2 ,

max . nc = 10 ,

method = " c o m p l e t e " ,

i n d e x =" a l l " )

nbdf2175

f v i z _ n b c l u s t ( nbdf2 ) + theme_minimal ( )

dev . o f f ( ) # d e l e t e t h e ’# ’ s i g n whenever

# you want t o c l e a n t h e p l o t s c r e e n

6


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


d i s t d f 2 . r e s <− d i s t ( df2 ,

method = " e u c l i d e a n " )180

h ca d f 2 <− h c l u s t ( d i s t d f 2 . r e s ,

method = " c o m p l e t e " )

p l o t ( hcadf2 ,

hang = −1) # dendogram v i s

r e c t . h c l u s t ( hcadf2 ,185

k = 3 ,

b o r d e r = 2 : 4 ) # dendogram v i s wi th g r o u p i n g

#### r o t a t i n g t h e p l o t

190

#### u s i n g ape

# l o a d package ape ; remember t o i n s t a l l i t : i n s t a l l . p a c k a g e s ( ’ ape ’ )

i n s t a l l . p a c k a g e s ( " ape " )

l i b r a r y ( ape )

p l o t ( a s . phy lo ( h c a d f 2 ) ,195

cex = 0 . 9 ,

l a b e l . o f f s e t = 1 ,

t y p e = " u n r o o t e d " )

p l o t ( a s . phy lo ( h c a d f 2 ) ,200

cex = 0 . 9 ,

l a b e l . o f f s e t = 1 )

4.2.3 Principal component analysis (PCA)

The PCA is applied using R base function and visualized using factominer and factoextra

packages. The following is the code.205

df <− as . d a t a . f rame ( r e a d . csv ( " data_smg . csv " ) ) # l o a d i n g as d a t a f rame

head ( d f ) # c h e c k i n g h e a d e r

i s . na ( d f ) # c h e c k i n g NAs i n d f

df2 <− df [ c ( 2 , 5 : 1 8 ) ] # s u b s e t t i n g df , e x c l u d e v a r wi th NAs

head ( df2 )210

i s . na ( d f2 ) # c h e c k i n g NAs i n df2


i s . numer ic ( d f2 ) # c h e c k i n g d a t a t y p e i n df2

7


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


rownames ( df2 ) <− d f 2 $ l o c a t i o n # s e t t i n g c o l l o c a t i o n as row names


i n s t a l l . p a c k a g e s ( " FactoMineR " )

l i b r a r y ( " FactoMineR " )

l i b r a r y ( f a c t o e x t r a )

r e s . pca <− PCA( df2 , g raph = FALSE)220

e i g e n v a l u e s <− r e s . p c a $ e i g

head ( e i g e n v a l u e s [ , 1 : 2 ] )

b a r p l o t ( e i g e n v a l u e s [ , 2 ] , names . a r g =1: nrow ( e i g e n v a l u e s ) ,

main = " V a r i a n c e s " ,

x l a b = " P r i n c i p a l Components " ,225

y l a b = " P e r c e n t a g e o f v a r i a n c e s " ,

c o l =" s t e e l b l u e " )

# Add c o n n e c t e d l i n e segmen t s t o t h e p l o t

l i n e s ( x = 1 : nrow ( e i g e n v a l u e s ) , e i g e n v a l u e s [ , 2 ] ,

t y p e =" b " , pch =19 , c o l = " r e d " )230

r e s . p c a $ v a r $ c o n t r i b

f v i z _ p c a _ v a r ( r e s . pca )

f v i z _ p c a _ v a r ( r e s . pca , c o l . v a r =" s t e e l b l u e " )+

theme_minimal ( )235

r e s . p c a $ i n d $ c o n t r i b

p l o t ( r e s . pca , c h o i x = " i n d " )

f v i z _ p c a _ b i p l o t ( r e s . pca , geom = " t e x t " )240

5 Conclusions

The present study integrates geological, hydrogeological data, and statistical analysis to construct

a hydrogeological model of the aquifer system in Semarang. The statistical treatment shows a con-

sistent pattern of anomalous setting at well point 37 (University Sultan Agung 2/Unisula-2). The

anomaly needs more in depth analysis to understand the underlying processes in the groundwater245

flow.

This paper is one of our preliminary example of data paper in Indonesia. Hopefully this can trigger

more data papers to endorse open science in our country.

8


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


Acknowledgements. The authors are thankful to the Department of Energy and Resources of Central Java

Province and Geological Agency in Bandung for providing hydrogeological data. Hopefully this paper will250

initiate a mass movement on open government data and data reuse in Indonesia.

9


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


References

Adams, S., Titus, R., Pietersen, K., Tredoux, G., and Harris, C.: Hydrochemical characteristics of aquifers near

Sutherland in the Western Karoo, South Africa, Journal of Hydrology, 241, 91–103, 2001.

Ayenew, T., Fikre, S., Wisotzky, F., Demlie, M., and Wohnlich, S.: Hierarchical cluster analysis of hydrochem-255

ical data as a tool for assessing the evolution and dynamics of groundwater across the Ethiopian rift, Interna-

tional journal of physical sciences, 4, 76–90, http://www.academicjournals.org/journal/IJPS/article-abstract/

D64DFAE18634, 2009.

BSN: Standard tests for water sample, Tech. rep., National Board for Standards, http://sisni.bsn.go.id/index.

php/sni_main/sni/detail_sni/7689, 2012.260

Coghlan, A.: Little Book of R for Multivariate Analysis! — Multivariate Analysis 0.1 documentation, Wellcome

Trust Sanger Institute, Cambridge, U.K., https://little-book-of-r-for-multivariate-analysis.readthedocs.io/en/

latest/, affiliation: Wellcome Trust Sanger Institute, Cambridge, U.K., 2009.

Deon, F., Förster, H.-J., Brehme, M., Wiegand, B., Scheytt, T., Moeck, I., Jaya, M., and Putriatni, D.: Geochem-

ical/hydrochemical evaluation of the geothermal potential of the Lamongan volcanic field (Eastern Java,265

Indonesia), Geothermal Energy, 3, 1–21, doi:10.1186/s40517-015-0040-6, 2015.

Herdianita, N. R., Julinawati, T., and Amorita, I. E.: Hydrogeochemistry of Thermal Water from Surface

Manifestation at Gunung Ciremai and Its Surrounding, Cirebon, West Java–Indonesia, in: Proceedings

World Geothermal Congress 2010, http://www.geothermal-energy.org/pdf/IGAstandard/WGC/2010/1476.

pdf, 2010.270

Irawan, D. E. and Putranto, T. A.: Dataset: hydrochemical assessment of Semarang area, Indonesia,

doi:10.5281/zenodo.57293, http://dx.doi.org/10.5281/zenodo.57293, 2016.

Irawan, D. E., Puradimaja, D. J., Notosiswoyo, S., and Soemintadiredja, P.: Hydrogeochemistry of volcanic

hydrogeology based on cluster analysis of Mount Ciremai, West Java, Indonesia, Journal of hydrology, 376,

221–234, http://www.sciencedirect.com/science/article/pii/S002216940900434X, 2009.275

Kassambara, A.: ggcorrplot: Visualization of a Correlation Matrix using ’ggplot2’, https://CRAN.R-project.

org/package=ggcorrplot, r package version 0.1.1, 2016.

Kassambara, A. and Mundt, F.: factoextra: Extract and Visualize the Results of Multivariate Data Analyses,

https://CRAN.R-project.org/package=factoextra, r package version 1.0.3, 2016.

King, A. C., Raiber, M., and Cox, M. E.: Multivariate statistical analysis of hydrochemical data to assess280

alluvial aquifer–stream connectivity during drought and flood: Cressbrook Creek, southeast Queensland,

Australia, Hydrogeology Journal, 22, 481–500, doi:10.1007/s10040-013-1057-1, http://link.springer.com/

10.1007/s10040-013-1057-1, 2014.

Lê, S., Josse, J., and Husson, F.: FactoMineR: A Package for Multivariate Analysis, Journal of Statistical Soft-

ware, 25, 1–18, doi:10.18637/jss.v025.i01, 2008.285

Maechler, M., Rousseeuw, P., Struyf, A., Hubert, M., and Hornik, K.: cluster: Cluster Analysis Basics and

Extensions, r package version 2.0.4 — For new features, see the ’Changelog’ file (in the package source),

2016.

Paradis, E., Claude, J., and Strimmer, K.: APE: analyses of phylogenetics and evolution in R language, Bioin-

formatics, 20, 289–290, 2004.290

10


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


Putranto, T. and Rüde, T.: Hydrogeological Model of an Urban City in a Coastal Area, Case study: Se-

marang, Indonesia, Indonesian Journal on Geoscience, 3, 17–27, https://ijog.geologi.esdm.go.id/index.php/

IJOG/article/view/227, 2016.

R Core Team: R: A Language and Environment for Statistical Computing, R Foundation for Statistical Com-

puting, Vienna, Austria, https://www.R-project.org/, 2016.295

Rahmawati, N. and Marfai, M.: Salinity Pattern in Semarang Coastal City: An Overview, Indonesian Jour-

nal of Geosciences, 8, doi:10.17014/ijog.3.1.17-27, https://ijog.geologi.esdm.go.id/index.php/IJOG/article/

view/160, 2013.

Wickham, H.: ggplot2: Elegant Graphics for Data Analysis, Springer-Verlag New York, http://ggplot2.org,

2009.300

Wickham, H. and Francois, R.: dplyr: A Grammar of Data Manipulation, https://CRAN.R-project.org/package=

dplyr, r package version 0.5.0, 2016.

Wilkinson, D. J.: Multivariate Data Analysis using R: a course notes, Tech. rep., https://www.staff.ncl.ac.uk/d.

j.wilkinson/teaching/mas8381/notes14.pdf, 2014.

11


Ope

n A

cces

s Earth System

Science

DataD

iscussio

ns


Published: 28 July 2016 Author(s) 2016. CC-BY 3.0 License ... · 2 Faculty of Engineering, Universitas Diponegoro, Jalan Prof. H. Soedarto, SH, Tembalang, Kota Semarang - 50275, Indonesia

Documents