International Journal Information System and Computer Full Paper

eISSN : 2598-246X
Science (IJISCS)
pISSN : 2598-0793

COMPARING TOOLS PROVIDED BY PYTHON

AND R FOR EXPLORATORY DATA ANALYSIS
Mahathir Rahmany1, Abdullah Mohd Zin2, Elankovan A.
Sundararajan3 *Corresponding author
1,2,3Centre for Software Technology and Management,
[email protected]
Faculty of Information Science and Technology, Universiti
Kebangsaan Malaysia (UKM),
Bangi, Selangor, 43600, Malaysia,
E-Mail: [email protected],
[email protected], [email protected]

Article history: Abstract

Received: 18 August 2020 To uncover the insight behind the data, the comprehensive
Revised: 14 October 2020 analysis is needed. Exploratory Data Analysis (EDA) is one of
Accepted: 24 October 2020 practical data analysis that will guide how to reveal any
hidden information in the data. By doing EDA, any pattern and
issue in the data will be seen and eventually will lead
Keywords:
hypothesis. To do EDA, beside any basic statistic is needed, a
exploratory data analysis,
good tool to simplify the analysis is also a consideration. Python
EDA,
and R as a famous programming language in data science
Python,
world provide method to implement that analysis. This paper
R,
will show how to perform EDA by utilizing the Python and R
statistics.
programming.

1.0 INTRODUCTION
Data cleaning and preparation process took almost 80% from the whole process in
statistical analysis [1]. In doing so, it needs a strategy to simplify this profoundly difficult process.
One of well known strategy is Exploratory Data Analysis (EDA) [2], [3]. EDA is very decisive
analysis in Data Science world. It is initial step [4]–[6] to thoroughly grasp the data by find out
any powerful hints such as suspected pattern [7], [8], anomaly [7], and also testing hypothesis
[9] and checking assumption [10] by using descriptive statistics and graphical tool [11], [12].
EDA is not one stop process but it is multiple process [13] until the data is increasingly clear what
is about [7]. Knowing these any disturbance or pattern in data beforehand, ultimately would
lead an actionable insight.
The main purpose of EDA is striving to derive insight from the data. To that end, EDA provide
variety of means to achieve that goal. On top of this, EDA is introduced not for statistician only
but for non statisticians can perform EDA too as the nature of EDA is seek and trial [7]. Even the
concept of EDA has been used years ago but the research about or the research utilize the
EDA concept still ever-expanding [7], [14]. Those things show how imperative the EDA.
Before EDA was introduced by Tukey [4], [13], [15], [16], Confirmatory Data Analysis (CDA)
is more well known compare to EDA. However, EDA and CDA are include in practical data
analysis [17]. Different from EDA which is to generate hypothesis as the result from the analysis,
CDA is more to test the already generated by the EDA before [7], [18], [19]. But both of them
are performed after pre-processing analysis [20]. So in this case it shows that EDA and CDA is
complementary each other [21].
There are 6 six steps to perform EDA [22]. Those are (i) Go through the attributes; (ii) Analyze
the univariate data; (iii) Grasp linkage among attributes; (iv) Observe suspected pattern such
as missing value and non-uniform value; (v) detect outlier; and (vi) feature engineering.

IJISCS | 131
 i. Go through the attribute.
Understanding the attribute on the data is very important task. By knowing the attribute
of the data for example which column is nominal, ordinal, discrete or continuous, the
type of statistical analysis that will be performed on that column can easily to be decided
[23]. In addition by going through the attribute, the data become more familiar [24].

ii. Analyze the univariate data.

Analyze for every single variable of data to understand the form of data. For example
analyze the centrality or dispersion of data [22]. By analyzing each variable, any
suspected pattern on the data can be spotted easily. The figure 1 below showed the
classification of descriptive statistics, where it consists with central tendency and
dispersion.

iii. Grasp linkage among attributes

Determine if each attribute have linkages between one another. The linkage is could be
correlation, covariance, and scatterplot.
Central tendency Dispersion

Mode Geometric Standard Deviation

Median Range

Mean Standard Deviation

Variance

Figure 1: The example of Central Tendency and Dispersion [25], [26]

iv. Observe suspected pattern

Suspected pattern might be missed value or any pattern that look unusual or uncommon.
Any action must be applied to the data in order it would not give any effect to the result
analysis.

v. Detect outlier
Outlier is any data point which clearly separated from other data [27]. The outlier can
influence the analysis result conclusion and as a result can generate inaccurate finding
[28].

vi. Feature engineering

Feature engineering is to prepare the raw data to be applied in machine learning model
by utilize the feature of the data [29].

There are many tools nowadays to perform EDA concept. They belong to the programming
and nonprogramming tools. Some well-known programming tools are Python and R. Those two
languages have been widely used be cutting edge programming language by researcher
and scientist to do data analysis especially EDA as their prominence in the availability tool.

IJISCS | 132
 Figure 2: Technique in EDA: Descriptive and Graphical [2], [12], [16], [30]–[33]

So, in this paper will compare these two programing tools for the purpose of EDA.

2.0 TECHNIQUES IN EXPLORATORY DATA ANALYSIS

There are three types of data: (i) univariate; (ii) Bivariate; and (iii) Multivariate. Each those
type of data has a different technique to applied EDA: Descriptive Statistical and graphical
techniques [4]. The most common techniques is graphical [8], [34].
Figure 2 denote some type of analyzing and also type of graph or plot that commonly used
for performing the EDA.
Graphical how the independent variable can affect the
Graphical in EDA means how to visualize dependent variable [38].
the data [4] in certain ways in order to get
actionable insight. From figure 2, it shows that Multivariate
every type of data has its own way to visualize Multivariate is statistical method that can
the data even with the same way, such as like perform analytics on more than one
Univariate and Bivariate data share same relationship outright [39]. As literally
type of visualization which is Histogram and multivariate data consist many variables so it
Boxplots. become arduous to analyze [40].

Descriptive Statistics Central tendency

Descriptive statistics is the process to The computation of centralized of data
summarize [26] the data, be it qualitative or by evaluated the scattered of data [41], [42].
quantitative [35]. It is also an analysis to assess Mean, median, and mode are the example
that the data is ready for the next analysis of central tendency analysis.
[36]. Dispersion
To estimate the dispersion of data [43].
Univariate The example of this kind of analysis are
Univariate is statistical method in analyzing variance, range, and standard deviation.
the data with exclusively single or individual
variable [12], [37], for example variable height Mean
only or variable weight only. As it is strictly The calculation of average on the data.
single variable so it is very simple to analyze
[34]. Median
The middle value of data when it is arranged
Bivariate from smallest to largest. The median is also
Bivariate is statistical method in analyzing known as 50th centile or Q2 [44].
two variables [30] only, for example height
and weight or education level and salary. The
purpose of bivariate analysis is to perceive
IJISCS | 133
Mode Histogram
Mode is the value that most often appears in Histogram is a graph to describe the
a set of data. frequency of continues variable [43], [51].

Standard deviation Stem plots

Standard deviation (SD) is measurement the Stem plots or also known as stem and leaf plot
dispersion of data from the mean [31], [45]. is a graph by showing the whole data values
distribution [37], [53].
Range
Range is one of measurement for dispersion Box plot
[33] to observe the dispersion from the lowest Box plot can describe either univariate or
value to highest value in a dataset. bivariate analysis. Box plot which is very good
on spotting the outlier [51], is a graph to
Variance describe the variable distribution and to
Variance is the distance between the single determinant any values that locate outside of
data from the mean [46], [47]. So we could others common value [51] or commonly
say that the variance is the deviation of mean known as outlier.
squared [48].
2D line plot
Cross-tabulation The graph that contain y and x axis where to
Cross tabulation is to perceive how two or describe array value on y axis and at x axis is
more categorical variable interconnected equal intervals [37].
[49], [50].
Scatterplot
Covariance The graph to visualize the relationship of two
Covariance is a type of descriptive statistics to variables [54]
see the further of the interdependence of two
variables [51], [52], Bar plot
To visualize the categorical data with bar
Correlation shape by directly proportional with value of
Correlation is to observe how the relationship the data.
of two variable associated [36], [51].

3.0 TOOLS FOR EDA

Python and R are most common programming tools to perform EDA [6], [7], [55]. R and
python are some of a very well known programming for data science [56]–[58].

3.1 Python
Python is one of common and famous programming language in Data Science. Even it is
built not only for data analysis, but it has powerful tool (in Python world, it is known as library) to
perform data analysis.
Some common libraries that is used to applied EDA in Python are Pandas [59]–[61], Numpy
[60], [62] and Matplotlib [63]. Pandas is a library in Python that is used for dealing and
processing the data [59], [61]. On other hand Numpy is a library to compute array based
numerical application [62]. Similarly with matplotlib, the library that is focusing on making graph
with many optional feature [63].
Beside those three library above there are also some libraries that really useful for EDA such
as Seaborn [64], [65] and Bokeh [65].

3.2 R
R is a programming language that focus on statistics and mathematics. In term of EDA,
which is belong to statistical concept, R has provided some package to deal with EDA easily.
There are some options to perform EDA in R such as using generic package or utilizing the
package that is aimed for EDA only. The generic package that is commonly used in
visualization EDA such as GGPLOT and Plotly. In other side there are some package that is built
only to perform EDA, for example SmartEDA.

IJISCS | 134
 The differentiate between generic package and built-in package in exploratory data
analysis in R is by using the built-in package, it provides command to perform exploratory data
analysis to guide on the analysis, but not for generic one. In generic package, to perform EDA
is by using common command.

3.3 Non-Programming Tools

Beside the fully programming tool as mention above, there are also some non programming
tool to perform EDA such as PowerBI, Tableau, Microsoft Excel, Google Data Studio and Google
Sheet.
PowerBI is a business intelligence tool that belongs to Microsoft. Beside PowerBI, Microsoft
also has another tool for analyzing data which is Microsoft Excel. Beside those tools, Google
also has non-programming tools to perform EDA which are known as Google Data Studio and
Google Sheet. Google Data Studio and Google Sheet are an analysis tool that is run by the
cloud.

4.0 COMPARISON
Comparison between the use of Python and R in carrying out EDA is done through a case
study.

4.1 The case study

To show comparison to use Python and R for EDA, the example questionnaire survey data
made. The data contains 16 attributes and 300 respondents. So, it means this data has 16
columns and 300 rows.
The classification of those 16 attributes are age, status and weight are numerical data;
gender, status, and disease are categorical data. The Q1 until Q10 are Likert Scale data.

4.2 Using Python

a. Go through attribute.
There are two ways to go through the attribute using python: (i) describe, and (ii) head.
Both will show the attribute name.
The figure 3 below show how to go through attribute using command head in python.

Figure 3: Attribute name

b. Analyze the univariate data

The figure 4 below show how to analyze the univariate data in python. It clearly shows
that some descriptive statistic is well compute there.

Figure 4: Analyze the univariate data

c. Grasp linkage among attributes
IJISCS | 135
 The figure 5 below show the visualization correlation among the attribute. In this case the
library matplotlib is used to draw the figure.

Figure 5: Correlation visualization

d. Observe suspected pattern

The figure 6 show that there is some attribute in dataset does not appropriate for next
computation. For example, the Q1 attribute is supposed to be a factor, not a numeric because
it is a likert scale.

Figure 6: Data type

e. Detect outlier
From the figure 7, it shows that there does not have any outlier on the data.

IJISCS | 136
 Figure 7: Detect outlier

4.3 Using R
a. Go through attribute.
There are two ways to go through the attribute: (i) head(), and (ii) summary(). Head() will
show top six of the data include the name of the attribute data as shown on figure 8. In other
side the summary() will show more detail about the data for example the minimal value, the
maximal value, first quarter of the data, median, mean and third quarter of the.

Figure 8: head ()

From the figure 8 above we can see the dataset attribute name such as: Gender, Age,
Weight, Height, Disease, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8. Q9 and Q10.

b. Analyze the univariate data

In this example we analyze the univariate by showing the histogram of attribute age. The
command to draw this graph is hist(dataset$attributename)

Figure 9: The histogram of attribute Age

From the figure 9 above we can see that the majority age of respondent is between 35 to
40. In the other side the age between 65 to 70 the minority age of the respondent.

c. Grasp linkage among attributes

We will see is there any relationship between the likert scale value, Q1 until Q10. From the
figure 10 shows that the dispersion of the answer from the respondent is same.

IJISCS | 137
 Figure 10: Bar plot for likert scale

d. Observe suspected pattern

In step four we want to check the data type of each attribute. If the data type is
appropriate with the assign data.

Figure 11: Data type in data frame.

From the figure 11 above it shows that the data type for each attribute is appropriate as its
assign.

e. Detect outlier
To detect outlier, we use syntax boxplot in R. From the figure 12 below, it shows there is no
outlier in attribute weight.

Figure 12: Detect outlier for attribute weight

IJISCS | 138
5.0 RESULT AND DISCUSSION
5.1 Result
From the result of case study above, it shows that both Python and R programming
language can perform the EDA smoothly. This is because both of those programming
languages has provided ready-to-use syntax in their programming. For R maybe this thing is not
weird because R developed to focus on statistical analysis, so it has provided any kind tool and
syntax to support statistical analysis and no exception for EDA.
For python as a universal programming language, also has gear itself with statistical tool to
support the EDA. Even though it does not have any special built-in package to perform EDA,
as what R has, but its common package for statistical analysis is good enough to carry out EDA,
for example Pandas, Numpy and Matplotlib.

5.2 Discussion
EDA is not a fixed analysis which mean that by doing the same analysis method will guide
any insight in data. But EDA is iterative analysis where any possibility analysis must be performed
on the data until what the data is saying is fully satisfied and understandable.
Therefore, what is shows in the Case Study section above will not fully demonstrate to find
any diamond on the data. That section only want to point out that Python and R is an option
that can be used to EDA beside any other options where maybe can fully accomplish the EDA
for example by using Microsoft Excel, Tableau, or SPSS.
As a nature of EDA where the analysis should be done iteratively in order to find any insight
information in the data, using the generic package or library is a good deal on choosing any
probability that may suit to the analysis. But for the built-in package or library tied to the
reserved command or analysis that has been provided. Despite that, harnessing the built-in
package or library is good start to open idea on what the data talking about, then going deep
into data by using the generic package or library.

6.0 CONCLUSION
EDA is an analysis to find any insight in data that will lead to hypothesis. EDA is repetitive
analysis to find any information inside the data. The analysis can be done in any means without
any terms that must be obeyed. However to simplify the analysis, there are some steps [22] that
can be followed: (i) Go through attribute, (ii) Analyze the univariate data, (iii) Grasp linkage
among attributes, (iv) Observe suspected pattern, (v) Detect outlier, and (vi) Feature
engineering.
There are two type of EDA, graphical and descriptive statistics. The graphical is the way to
explore data by visualizing it. In other side, to understand the data by quantitative computation
is descriptive statistics. Python and R provide solution to perform EDA. With the complete syntax,
exploring the data with them is very easy.

REFERENCES
[1] M. Huebner, W. Vach, dan S. [le Cessie], “A systematic approach to initial data analysis
is good research practice,” J. Thorac. Cardiovasc. Surg., vol. 151, no. 1, hal. 25–27, 2016.
[2] M. Staniak dan P. Biecek, “The Landscape of R Packages for Automated Exploratory
Data Analysis,” arXiv e-prints, hal. arXiv:1904.02101, Mar 2019.
[3] K. A. Monsen, “Exploratory Data Analysis,” in Intervention Effectiveness Research: Quality
Improvement and Program Evaluation, Cham: Springer International Publishing, 2018,
hal. 77–85.
[4] S. Putatunda, K. Rama, D. Ubrangala, dan R. Kondapalli, “SmartEDA: An R Package for
Automated Exploratory Data Analysis,” arXiv Prepr. arXiv1903.04754, 2019.
[5] C. M. Carbery, R. Woods, dan A. H. Marshall, “A New Data Analytics Framework
Emphasising Pre-processing in Learning AI Models for Complex Manufacturing Systems,”
in Intelligent Computing and Internet of Things, 2018, hal. 169–179.
[6] G. L. Taboada, I. Seruca, C. Sousa, dan Á. Pereira, “Exploratory Data Analysis and Data
Envelopment Analysis of Construction and Demolition Waste Management in the
European Economic Area,” Sustainability, vol. 12, no. 12, hal. 4995, 2020.
[7] A. Bezerra, I. Silva, L. A. Guedes, D. Silva, G. Leitão, dan K. Saito, “Extracting Value from

IJISCS | 139
 Industrial Alarms and Events: A Data-Driven Approach Based on Exploratory Data
Analysis,” Sensors, vol. 19, no. 12, hal. 2772, Jun 2019.
[8] “Understanding Clinical Data using Exploratory Analysis,” Int. J. Recent Technol. Eng., vol.
8, no. 5, hal. 5434–5437, Jan 2020.
[9] Z. Jones dan F. Linder, “Exploratory data analysis using random forests,” in Prepared for
the 73rd annual MPSA conference, 2015.
[10] V. Indhumathi dan N. Dharani, “Estimation of Data Analysis In R to Predict Diabetes,” Int.
J. Emerg. Technol. Innov. Eng., vol. 6, no. 01, 2020.
[11] E. Camizuli dan E. J. Carranza, “Exploratory Data Analysis (EDA),” in The Encyclopedia of
Archaeological Sciences, American Cancer Society, 2018, hal. 1–7.
[12] S. M. Thaung, H. M. Tun, K. K. K. Win, M. M. Than, dan A. S. S. Phyo, “Exploratory data
analysis based on remote health care monitoring system by using IoT,” Communications,
vol. 8, no. 1, hal. 1–8, 2020.
[13] Y. Mao dan others, “Data Visualization in Exploratory Data Analysis: An Overview of
Methods and Technologies,” 2015.
[14] S. Kaleru dan S. R. Dhanikonda, “Exploratory Data Analysis and Latent Dirichlet Allocation
on Yelp Database,” Int. J. Appl. Eng. Res., vol. 13, no. 21, hal. 15035–15039, 2018.
[15] J. W. Tukey, Exploratory data analysis, vol. 2. Reading, MA, 1977.
[16] K. Wongsuphasawat, Y. Liu, dan J. Heer, “Goals, Process, and Challenges of Exploratory
Data Analysis: An Interview Study,” arXiv Prepr. arXiv1911.00568, 2019.
[17] S. Thiprungsri, M. A. Vasarhelyi, A. Kogan, M. Alles, dan J. J. Ye, “Cluster analysis for
anomaly detection in accounting,” in Rutgers Studies in Accounting Analytics: Audit
Analytics in the Financial Industry (Rutgers Studies in Accounting Analytics), Emerald
Publishing Limited, 2019, hal. 87–110.
[18] T. T. Allen, Z. Sui, dan K. Akbari, “Exploratory text data analysis for quality hypothesis
generation,” Qual. Eng., vol. 30, no. 4, hal. 701–712, 2018.
[19] Z. Maznah, M. Halimah, M. Shitan, P. K. Karmokar, dan S. Najwa, “Prediction of
Hexaconazole Concentration in the Top Most Layer of Oil Palm Plantation Soil Using
Exploratory Data Analysis ({EDA}),” {PLOS} {ONE}, vol. 12, no. 1, hal. e0166203, Jan 2017.
[20] X. Ma et al., “Using Visual Exploratory Data Analysis to Facilitate Collaboration and
Hypothesis Generation in Cross-Disciplinary Research,” ISPRS Int. J. Geo-Information, vol.
6, no. 11, 2017.
[21] C. M. Igwenagu, “EXPLORATORY DATA ANALYSIS AND MULTIVARIATE STRATEGIES FOR
REVEALING MULTIVARIATE STRUCTURES IN CLIMATE DATA,” 2016.
[22] A. Ghosh, M. Nashaat, J. Miller, S. Quader, dan C. Marston, “A comprehensive review of
tools for exploratory analysis of tabular industrial datasets,” Vis. Informatics, vol. 2, no. 4,
hal. 235–253, 2018.
[23] J. R. Dettori dan D. C. Norvell, “The Anatomy of Data,” Glob. Spine J., vol. 8, no. 3, hal.
311–313, Jan 2018.
[24] V. Cox, “Exploratory Data Analysis,” in Translating Statistics to Make Decisions : A Guide
for the Non-Statistician, Berkeley, CA: Apress, 2017, hal. 47–74.
[25] S. Deshpande, N. J. Gogtay, dan U. M. Thatte, “Measures of central tendency and
dispersion,” J. Assoc. Physicians India, vol. 64, hal. 64–66, 2016.
[26] E. G. M. Hui, “Descriptive Statistics,” in Learn R for Applied Statistics: With Data
Visualizations, Regressions, and Statistics, Berkeley, CA: Apress, 2019, hal. 87–127.
[27] C. C. Aggarwal, “An Introduction to Outlier Analysis,” in Outlier Analysis, Cham: Springer
International Publishing, 2017, hal. 1–34.
[28] M. Z. Iqbal, S. Habib, M. I. Khan, dan M. Kashif, “COMPARISON OF DIFFERENT TECHNIQUES
FOR DETECTION OF OUTLIERS IN CASE OF MULTIVARIATE DATA,” Pak. J. Agri. Sci, vol. 57,
no. 3, hal. 865–869, 2020.
[29] A. Zheng dan A. Casari, Feature Engineering for Machine Learning: Principles and
Techniques for Data Scientists. O’Reilly Media, 2018.
[30] A. G. Bronevich dan J. V. de Oliveira, “On the model updating operators in univariate
estimation of distribution algorithms,” Nat. Comput., vol. 15, no. 2, hal. 335–354, Mei 2015.
[31] R. W. Cooksey, “Descriptive Statistics for Summarising Data,” in Illustrating Statistical
Procedures: Finding Meaning in Quantitative Data, Singapore: Springer Singapore, 2020,
hal. 61–139.

IJISCS | 140
[32] M. Allen, The {SAGE} Encyclopedia of Communication Research Methods. {SAGE}
Publications, Inc, 2017.
[33] E. Mooi, M. Sarstedt, dan I. Mooi-Reci, “Descriptive Statistics,” in Springer Texts in Business
and Economics, Springer Singapore, 2017, hal. 95–152.
[34] N. Iftikhar, T. Baattrup-Andersen, F. Nordbjerg, E. Bobolea, dan P.-B. Radu, “Data
Analytics for Smart Manufacturing: A Case Study,” in Proceedings of the 8th International
Conference on Data Science, Technology and Applications, 2019.
[35] S. K. Sharma, T. Kanchan, dan K. Krishan, “Descriptive Statistics,” in The Encyclopedia of
Archaeological Sciences, American Cancer Society, 2018, hal. 1–8.
[36] T. C. Guetterman, “Basics of statistics for primary care research,” Fam. Med. Community
Heal., vol. 7, no. 2, hal. e000067, Mar 2019.
[37] M. Komorowski, D. C. Marshall, J. D. Salciccioli, dan Y. Crutain, “Exploratory Data
Analysis,” in Secondary Analysis of Electronic Health Records, Cham: Springer
International Publishing, 2016, hal. 185–203.
[38] A. Bertani, G. Di Paola, E. Russo, dan F. Tuzzolino, “How to describe bivariate data,” J.
Thorac. Dis., vol. 10, no. 2, hal. 1133–1137, Feb 2018.
[39] S. McQuitty, “The Purposes of Multivariate Data Analysis Methods: an Applied
Commentary,” J. African Bus., vol. 19, no. 1, hal. 124–142, 2018.
[40] X. He, Y. Tao, Q. Wang, dan H. Lin, “A co-analysis framework for exploring multivariate
scientific data,” Vis. Informatics, vol. 2, no. 4, hal. 254–263, 2018.
[41] M. A. Islam dan A. Al-Shiha, “Basic Summary Statistics,” in Foundations of Biostatistics,
Singapore: Springer Singapore, 2018, hal. 39–72.
[42] U. S. Ali, “A Case Study on Teaching of Fundamental Aspects of Central Tendency by
Using Classroom Activities at Secondary School Level, Karachi, Pakistan,” RADS J. Soc.
Sci. Bus. Manag., vol. 3, no. 2, hal. 41–56, 2016.
[43] A. Gupta, P. Mishra, C. Pandey, U. Singh, C. Sahu, dan A. Keshri, “Descriptive statistics
and normality tests for statistical data,” Ann. Card. Anaesth., vol. 22, no. 1, hal. 67, 2019.
[44] P. Shah et al., “Pancreatic Glucagon secretion is severely impaired and Somatostatin
secretion unchanged in patients with Hyperinsulinaemic Hypoglycaemia,” in 55th
Annual ESPE, 2016, vol. 86.
[45] O. O. Mosobalaje, O. D. Orodu, dan D. Ogbe, “Descriptive statistics and probability
distributions of volumetric parameters of a Nigerian heavy oil and bitumen deposit,” J.
Pet. Explor. Prod. Technol., vol. 9, no. 1, hal. 645–661, Jun 2018.
[46] F. Kaliyadan dan V. Kulkarni, “Types of variables, descriptive statistics, and sample size,”
Indian Dermatol. Online J., vol. 10, no. 1, hal. 82, 2019.
[47] Z. Ali dan Sb. Bhaskar, “Basic statistical tools in research and data analysis,” Indian J.
Anaesth., vol. 60, no. 9, hal. 662, 2016.
[48] L. Igual dan S. Seguí, “Descriptive Statistics,” in Introduction to Data Science: A Python
Approach to Concepts, Techniques and Applications, Cham: Springer International
Publishing, 2017, hal. 29–50.
[49] R. Ewing dan K. Park, Basic Quantitative Research Methods for Urban Planners. Taylor &
Francis, 2020.
[50] S. Ismail dan S. Ahmed, “Air Pollution, its Sources and Health Effects: A Case Study of
Delhi,” Res. J. Soc. Sci., vol. 9, no. 4, hal. 62–74, 2018.
[51] M. Sarstedt dan E. Mooi, “Descriptive Statistics,” in A Concise Guide to Market Research:
The Process, Data, and Methods Using IBM SPSS Statistics, Berlin, Heidelberg: Springer
Berlin Heidelberg, 2019, hal. 91–150.
[52] A. Indrabudiman, “Descriptive Analysis stock price with Zmijewski bankruptcy model to
total assets on stock prices,” Int. J. Sci. Res. Sci. Technol. Vol., vol. 3, 2017.
[53] S. Boslaugh, Encyclopedia of Epidemiology. {SAGE} Publications, Inc., 2008.
[54] N. Fitzallen, “Interpreting Association from Graphical Displays.,” Math. Educ. Res. Gr.
Australas., 2016.
[55] D. Borcard, F. Gillet, dan P. Legendre, “Exploratory Data Analysis,” in Numerical Ecology
with R, Cham: Springer International Publishing, 2018, hal. 11–34.
[56] M. R. M. Huddar dan R. V Kulkarni, “Role of R and Python in Data Science,” Res. JOURNEY,
hal. 32, 2018.
[57] S. K. A. Fahad dan A. E. Yahya, “Big Data Visualization: Allotting by R and Python with

IJISCS | 141
 GUI Tools,” in 2018 International Conference on Smart Computing and Electronic
Enterprise (ICSCEE), 2018, hal. 1–8.
[58] C. D. Larose dan D. T. Larose, Data Science Using Python and R. Wiley, 2019.
[59] J. Bernard, “Python Data Analysis with pandas,” in Python Recipes Handbook: A
Problem-Solution Approach, Berkeley, CA: Apress, 2016, hal. 37–48.
[60] I. Meniailov, K. Bazilevych, K. Fedulov, dan S. Goranina, “Using the K-means method for
diagnosing cancer stage using the Pandas library,” development, vol. 14, hal. 15, 2019.
[61] F. Nelli, “The pandas Library---An Introduction,” in Python Data Analytics: With Pandas,
NumPy, and Matplotlib, Berkeley, CA: Apress, 2018, hal. 87–139.
[62] M. Bauer dan M. Garland, “Legate NumPy: Accelerated and Distributed Array
Computing,” in Proceedings of the International Conference for High Performance
Computing, Networking, Storage and Analysis, 2019.
[63] J. Hunt, “Introduction to Matplotlib,” in Advanced Guide to Python 3 Programming,
Cham: Springer International Publishing, 2019, hal. 35–42.
[64] E. Bisong, “Matplotlib and Seaborn,” in Building Machine Learning and Deep Learning
Models on Google Cloud Platform: A Comprehensive Guide for Beginners, Berkeley, CA:
Apress, 2019, hal. 151–165.
[65] O. Embarak, “Data Visualization,” in Data Analysis and Visualization Using Python:
Analyze Data to Create Visualizations for BI Systems, Berkeley, CA: Apress, 2018, hal. 293–
342.

IJISCS | 142