An Empirical Study on How the Developers Discussed about Pandas Topics Sajib Kumar Saha Joy Farzad Ahmed Al Hasib Mahamud and Nibir_2

2025-04-30 1 0 2.4MB 15 页 10玖币

侵权投诉

An Empirical Study on How the Developers

Discussed about Pandas Topics

Sajib Kumar Saha Joy?, Farzad Ahmed?, Al Hasib Mahamud?, and Nibir

Chandra Mandal

Ahsanullah University of Science and Technology, Bangladesh

{joyjft, farzadahmed6}@gmail.com, {hasib, nibir}.cse@aust.edu

Abstract. Pandas is deﬁned as a software library which is used for data

analysis in Python programming language. As pandas is a fast, easy and

open source data analysis tool, it is rapidly used in diﬀerent software

engineering projects like software development, machine learning, com-

puter vision, natural language processing, robotics, and others. So a huge

interests are shown in software developers regarding pandas and a huge

number of discussions are now becoming dominant in online developer

forums, like Stack Overﬂow (SO). Such discussions can help to under-

stand the popularity of pandas library and also can help to understand

the importance, prevalence, diﬃculties of pandas topics. The main aim

of this research paper is to ﬁnd the popularity and diﬃculty of pandas

topics. For this regard, SO posts are collected which are related to pan-

das topic discussions. Topic modeling are done on the textual contents of

the posts. We found 26 topics which we further categorized into 5 board

categories. We observed that developers discuss variety of pandas topics

in SO related to error and excepting handling, visualization, External

support, dataframe, and optimization. In addition, a trend chart is gen-

erated according to the discussion of topics in a predeﬁned time series.

The ﬁnding of this paper can provide a path to help the developers, ed-

ucators and learners. For example, beginner developers can learn most

important topics in pandas which are essential for develop any model.

Educators can understand the topics which seem hard to learners and

can build diﬀerent tutorials which can make that pandas topic under-

standable. From this empirical study it is possible to understand the

preferences of developers in pandas topic by processing their SO posts.

Keywords: Pandas ·Stack Overﬂow ·Natural Language Processing ·

Empirical Software Engineering.

1 Introduction

Pandas is an open source library where python package oﬀers data manipulation

and analysis in python programming language. As pandas has high performance

and fast productivity for users, it’s data analysis capability is utilized in diﬀer-

ent sectors of computer science like data visualization, machine learning, data

?First Three authors contributed equally to this research.

arXiv:2210.03519v2 [cs.SE] 10 May 2023

2 Joy et al.

driven software engineering, computer vision, natural language processing etc.

Since 2008, the development of pandas has removed the distance between the

availability of data analysis tools [4]. Pandas is considered for most suitable op-

tion for data analysis tool as it is written in python programming language and

easy to understand for new beginner [5].

In the recent years, the utilization of pandas library is increased rapidly as pan-

das library has reduced the gap between scientiﬁc programming languages and

database languages [17]. Pandas library is utilized in most of the sectors like

machine learning, statistics, natural language processing, computer vision and

others. Moreover, pandas library is easy to understand for a beginner and it is

open source tool. For these reasons, most of the developers are now showing

interest to utilize pandas tools in their projects. For the development of pandas

library and it’s utilization, a factor is observed that the discussions regarding

pandas in online developers forums has increased, such as Stack Overﬂow (SO).

From analysing these post, several ﬁndings can be achieved related to pandas li-

brary like it’s popularity, diﬃculties, future scopes etc. To date, there are around

22.44 million questions are posted in SO [6].Several research works are conducted

based on SO posts in the ﬁeld of IoT [8], blockchain [9], microservices [10], soft-

ware engineering [8].Some research works are also done based on the function-

ality, popularity, scope to development of pandas library [11–13]. According to

the best of our knowledge, no research work is done based on the SO posts of

pandas library to ﬁnd the topics, popularity, scopes of pandas library.

In this research paper, total 236711 SO posts where user deﬁned tags are related

to pandas are analyzed to ﬁnd the topics of pandas library. For topic modeling,

Latent Dirichlet Allocation (LDA) is performed. Finally trend chart is generated

to ﬁnd the popularity’s of the topics according to the discussions of the software

development forums. In this empirical study, some major ﬁndings are shown.

Among of the major ﬁndings, ﬁrstly we have found the topics and then we have

categories pandas topics which are discussed most in the SO posts. According to

the ﬁndings of this paper, there are total twenty six topics and these twenty six

topics can be categorized into six categories. Among of the topics optimization

is the most popular topic though SQL queries and Matplotlib support are the

most diﬃcult topics as SQL queries is having the lowest score and Matplotlib

support is having the lowest accepted answer rate. Secondly, to make a closer

look of the topics and categories, a trend chart is generated from the time slot

July 2011 to February 2022. Some decline and arises are seen in the trend chart

of the topics but the total number of posts are increased gradually as the total

amount of pandas developers increased over time.

The next of the paper is organized in the following way: Section 2 discusses

the background studies of this paper. Methodology is described in Section 3

where data collection, topic modeling and topic naming process are answered.

Section 4 discusses implication of studies where several important expositions

are described. Section 5 describes threats to the validity of our result. Section 6

describes results of our study where section 7 answers the future scopes to work

from the result. Section 8 concludes the paper.

An Empirical Study on How the Developers Discussed about Pandas Topics 3

2 Background Studies

2.1 Stack Overﬂow

Stack Overﬂow (SO) is considered as a question and answering sites which has

become popular in recent times for software developers forum. There are some

sites for programmers where programmers can ask questions, can answer other’s

questions and can share ideas. SO is one of them [14].

2.2 Topic Modeling

Topic modeling can be deﬁned as unsupervised classiﬁcation method which

is similar to clustering data and it is usually used for ﬁnding groups of item

[16].Though topic modeling is mostly used for textual data, it can be used for

bioinformatics data, social science data and other source of data [10].

To apply topic modeling in the SO posts related to pandas, in this paper Latent

Dirichlet Allocation (LDA) is used. In LDA, each document is considered as

combination of topics and each topic is considered as combination of words [16].

Topic modeling is used in diﬀerent papers of various ﬁelds where SO posts are

utilized, specially in the aspects of software engineering development [8].

2.3 Pandas

Pandas is an open source data analysis tool which is built on Python program-

ming language and it is considered as fast, ﬂexible and easy tool [16]. Pandas

works in structured dataset and can be leveraged in social science, statistics,

ﬁnance and other ﬁelds [17]. Since the development of pandas in 2008, the main

aim of pandas library development is to remove the distance between scientiﬁc

programming languages and database languages [17].

2.4 Related Works

There are several works related to LDA Topic Modeling [8,8,8,18–23]. Uddin et

al. [8] have done an empirical study based on the IoT discussions of IoT topic on

SO posts. For this purpose, authors have gathered IoT posts from SO and have

leveraged LDA to perform topic modeling. Four research questions are answered

to ﬁnd the discussion, evolve, question types, popularity and diﬃclties of IoT

topics.

Stephen W. Thomas [18] has leveraged statistical topic modeling in mining soft-

ware repositories to analyze unstructured and unlabeled data and has found

structure in the textual repositories. Bavota et al. [19] have shown the opportu-

nities of Move Method refactoring and have removed Feature Envy bad smells

from source code and for this purpose Relational Topic Models are utilized.Nabil

et al. [20] have utilized Topic modeling in cloud computing and to discover eﬃ-

cient cloud services, LDA is leveraged. To ﬁnd the research trends, methodology

and ﬁelds of further research in blockchain technology, Shahid et al. [21] have

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

AnEmpiricalStudyonHowtheDevelopersDiscussedaboutPandasTopicsSajibKumarSahaJoy?,FarzadAhmed?,AlHasibMahamud?,andNibirChandraMandalAhsanullahUniversityofScienceandTechnology,Bangladesh{joyjft,farzadahmed6}@gmail.com,{hasib,nibir}.cse@aust.eduAbstract.Pandasisdenedasasoftwarelibrarywhichisusedfordataa...

展开>> 收起<<

An Empirical Study on How the Developers Discussed about Pandas Topics Sajib Kumar Saha Joy Farzad Ahmed Al Hasib Mahamud and Nibir_2.pdf

共15页,预览3页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

An Empirical Study on How the Developers Discussed about Pandas Topics Sajib Kumar Saha Joy Farzad Ahmed Al Hasib Mahamud and Nibir_2

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: