GUFBD / GUJS (2025) 15(1): 197-209 doi: 10.17714/gumusfenbil.

1600286 Research Article

Next-generation software development competencies: Identification of technical

Fatih GÜRCAN*1 , Cemal KÖSE2

• Received: 12.12.2024 • Accepted: 14.02.2025

Abstract
The software development industry is undergoing unprecedented growth and transformation, prompting a reevaluation of
the skills and competencies necessary for success in this dynamic landscape. This study investigates the rapidly evolving
skill requirements within the industry, driven by technological advancements. To achieve this, a Latent Dirichlet
Allocation (LDA) framework is employed, enabling the identification of key topics from a dataset derived from online
job postings. The analysis revealed 52 core topics pertinent to software development competencies. The findings reveal
a significant emphasis on both technical domain knowledge and programming skills, with particular attention to modern
programming languages such as Java, Python, and JavaScript. Moreover, non-technical skills, including communication,
teamwork, and critical thinking, are underscored as vital competencies in today’s collaborative software development
environments. These insights emphasize the necessity for software developers to cultivate a diverse skill set to adapt to
current and future industry demands. This work serves as a crucial reference for understanding the present and future skill
requirements in the software development field, providing valuable guidance for developers, employers, and educational
institutions.

Keywords: Competency analysis, Skill requirements, Software development skills, Topic modeling

Öz
Yazılım geliştirme sektörü önemli bir büyüme ve dönüşümden geçiyor ve bu dinamik ortamda başarı için gerekli
becerilerin ve yeterliliklerin yeniden değerlendirilmesi gerekiyor. Bu çalışma, teknolojik gelişmelerin yönlendirdiği
sektördeki hızla değişen beceri gereksinimlerini araştırıyor. Bunu başarmak için, çevrimiçi iş ilanlarından türetilen bir
veri setinden temel konuların belirlenmesini sağlayan bir Gizli Dirichlet Tahsisi (Latent Dirichlet Allocation) çerçevesi
kullanılıyor. Gerçekleştirilen analiz, yazılım geliştirme yetkinliklerine özgü 52 temel beceriyi ortaya çıkarmıştır.
Bulgular, Java, Python ve JavaScript gibi modern programlama dillerine özel dikkat gösterilerek hem teknik alan
bilgisine hem de programlama becerilerinin önemine vurgu yapmıştır. Dahası, iletişim, ekip çalışması ve eleştirel
düşünme gibi teknik olmayan beceriler, günümüzün işbirlikçi yazılım geliştirme ortamlarında çok önemli yetkinlikler
ortaya çıkmıştır. Bu içgörüler, yazılım geliştiricilerinin mevcut ve gelecekteki endüstri taleplerine uyum sağlamak için
çeşitli bir beceri seti geliştirmeleri gerekliliğini vurgulamaktadır. Bu çalışma yazılım geliştirme alanındaki mevcut ve
gelecekteki beceri gereksinimlerini anlamak için önemli bir referans görevi görerek geliştiriciler, işverenler ve eğitim
kurumları için değerli rehberlik sağlamaktadır.

Anahtar kelimeler: Beceri gereksinimleri, Konu modelleme, Yazılım geliştirme becerileri, Yetkinlik analizi

*Fatih GÜRCAN; fgurcan@[Link]

ISSN: 2146-538X e-ISSN: 2146-538X [Link]

1. Introduction

The software industry has undergone significant transformation in recent years due to rapidly evolving
technologies and increasing competition. Modern software development processes now require a
multidisciplinary skill set that extends beyond technical knowledge and expertise. In this dynamic
environment, software developers must not only enhance their technical proficiency but also develop
competencies in areas such as analytical thinking, problem-solving, teamwork, and communication. As
technology advances, the demand for software development professionals is rapidly changing, with evolving
needs for programming languages and software development tools (Aken et al., 2010; Debortoli et al., 2014;
Gurcan, 2023a). In this context, understanding the latest trends in the software industry and identifying the
skills and knowledge required to meet these demands is crucial for both employers shaping the sector and
developers advancing their careers (Gurcan & Cagiltay, 2019).

Existing studies indicate that skill requirements in the software development field are evolving rapidly, and
the competencies needed by software professionals continue to expand (Aken et al., 2010; Thomas et al., 2014;
Terblanche & Wongthongtham, 2015; Chen et al., 2016; Vayansky & Kumar, 2020). A series of studies have
highlighted that with the increasing pace of technological innovations, developers need more than just technical
knowledge; they also require soft skills such as analytical thinking, problem-solving, and teamwork (Aken et
al., 2010; Moreno et al., 2012; Debortoli et al., 2014; Gurcan & Cagiltay, 2019). Furthermore, as software
projects become more complex, multiple studies have addressed the need for developers to be proficient in
various programming languages and to quickly adapt to new tools (Moreno et al., 2012; Terblanche &
Wongthongtham, 2015; Gurcan & Cagiltay, 2019). Probabilistic topic modeling techniques are frequently
employed to analyze these dynamic industrial skill requirements and to identify the expertise demanded of
software developers by examining large datasets from online job postings (Gurcan & Kose, 2017; De Mauro
et al., 2018; Gurcan & Cagiltay, 2019). In these studies, methods such as Latent Dirichlet Allocation (LDA)
help uncover hidden topics in text, revealing which skills are in greater demand and which techniques are
gaining importance in the industry (Blei et al., 2003; Blei, 2012). Particularly, changes in next-generation
programming languages, software development tools, and methodologies can be better understood through
these analyses, aiding developers in planning their careers (Barua et al., 2014; De Mauro et al., 2018; Gurcan,
2023a).

This paper presents an analysis aimed at understanding the impact of technological transformation and
innovation in the software development industry on the workforce, specifically focusing on the current
demands for software development expertise. The primary goal of the study is to identify the expert domains,
technical skills, competencies, and the necessary tools that form the foundation of the software development
process in continually evolving environments. Accordingly, a comprehensive analysis was conducted on a
large dataset of online job postings for software developers using the probabilistic topic modeling method
based on LDA (Latent Dirichlet Allocation). Probabilistic topic modeling is a powerful tool for discovering
hidden semantic structures in large datasets. In this study, the text data gathered from online job postings was
analyzed using the LDA algorithm to uncover the most relevant knowledge and skills related to software
development expertise. The 52 core topics identified were categorized into domains of expertise and technical
knowledge requirements to highlight current trends in the software industry. Thus, the crucial knowledge and
skills for software development processes were classified into four main competency areas: technical
knowledge and skills (23%), software development and programming skills (52%), analytical and logical skills
(12%), and non-technical skills (13%).

The findings demonstrate that the knowledge and skills required for software developers span a broad
spectrum, with these trends becoming increasingly interdisciplinary. In particular, the importance of non-
technical skills such as problem-solving abilities, teamwork, communication skills, and leadership, alongside
technical proficiency, has been emphasized. In this context, software developers must not only possess coding
skills but also be capable of collaborating effectively, working within teams, and making strategic decisions.
The results presented in this paper indicate that the competencies required in software development processes
have become multifaceted, and the development of these skills is critical for success in the industry. The
findings provide a valuable guide for software professionals, helping them adapt to changes in the sector and
shape their career development accordingly. This study contributes to identifying new trends in the software
industry, enabling professionals to acquire the knowledge and skills needed for future demands. Additionally,

198
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

by offering a taxonomy of both technical and non-technical skills in demand, it serves as a useful resource for
educational institutions and employers.

2. Material and methods

The architecture of the study involves the stages of semantic analysis using an LDA-based topic modeling
approach on a text-based qualitative dataset. In the first stage, an experimental dataset was created using data
obtained from online job platforms. Subsequently, preprocessing steps were applied to the dataset to reduce
dimensionality and enhance the accuracy of the analysis. In the next stage, the qualitative dataset was converted
into a Document-Term Matrix (DTM) to enable numerical analyses. Following this step, a semantic analysis
based on an LDA-based probabilistic topic modeling approach was performed on the DTM to uncover latent
semantic structures. The topics derived from the experimental analysis were then discussed in terms of their
semantic relevance, and the experimental analysis was reconfigured with different parameters until the desired
level of semantic coherence was achieved. In the final stage, the results and interpretations derived from the
analysis were presented and discussed in light of similar studies.

2.1. Data collection

In today’s technological environment, online job platforms are widely used by both employers and employees
(Debortoli et al., 2014; Gurcan & Cagiltay, 2019). These platforms serve as essential communication tools
facilitating interactions between employers and job seekers (Gurcan & Cagiltay, 2019). As part of this study,
several online job platforms were examined to determine a suitable data source. Two important criteria were
considered in selecting the data source for the experimental analysis: first, the selected platform needed to be
focused heavily on software development roles, and second, it needed to be a global platform featuring job
postings for software development roles across different countries.

Considering these criteria, Stack Overflow’s job platform (Stack Overflow Developer Jobs) was chosen as the
data source (Stack Overflow, 2024). The job postings on Stack Overflow are specifically focused on software
development roles (developer jobs) (Barua et al., 2014; Stack Overflow, 2024). Besides being a significant
platform in the software development industry, Stack Overflow is also an important knowledge-sharing
platform where developers interact through a question-and-answer format (Gurcan, 2023a). In this respect,
Stack Overflow holds strategic importance for software developers (Barua et al., 2014; Stack Overflow, 2024).
The dataset created for this study consists of 4,128 job postings focused on software development. Each record
in the dataset typically includes information such as job title, date, location, job description, required
knowledge, skills, qualifications, and competencies.

2.2. Data preprocessing

Text preprocessing is a critical step that directly impacts the performance of text-based data analyses (Gurcan,
2023b). The preprocessing procedures applied in this study included tokenization, text cleaning, removal of
incomplete or irrelevant text entries, elimination of stop words, and frequency reduction, among other
sequential processes (Gurcan, 2023b; Wu et al., 2024). Depending on the type of data being processed and the
nature of the experimental analysis, preprocessing steps may vary. Given the high technical jargon content of
the texts in this study’s dataset, stemming was not applied to these texts (Chen et al., 2016).

Following these steps, terms with a frequency of fewer than five occurrences (i.e., words with negligible
semantic contribution) were removed from the word space, further reducing the dimensionality of the dataset
(Chen et al., 2016). This completed the preprocessing phase. The resulting word vectors for each text in the
dataset were then compiled into a DTM (Burkhardt & Kramer, 2019). Once the preprocessing phase was
completed, each text in the dataset was represented as a word vector. In the vector space model used to model
the distribution of terms across documents, all numerical analyses were performed on these word vectors (Feng
et al., 2022). To model the entire dataset, the term vectors for each document were combined into a single
DTM.

199
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

2.3. Topic modeling analysis using LDA

Probabilistic topic modeling is a generative and probabilistic approach used to model the semantic structure of
text documents within large collections and uncover latent semantic patterns. Text documents contain latent
semantic structures known as topics (Blei et al., 2003; Blei, 2012). Each topic is characterized by a probability
distribution over a fixed vocabulary of words. According to the probabilistic topic modeling approach, a text
document can contain multiple topics with varying proportions (Barua et al., 2014; Gurcan & Cagiltay, 2019).
The determination of these topic proportions forms the core of the probabilistic topic modeling approach.

In this study, we chose the Latent Dirichlet Allocation (LDA) method for topic modeling due to its proven
effectiveness in discovering latent topics within large text datasets. LDA is widely used in text mining and
provides a balance between interpretability and computational efficiency, making it suitable for our dataset
(Blei et al., 2003; Blei, 2012). Although deep learning-based methods such as BERT (Bidirectional Encoder
Representations from Transformers) and GPT (Generative Pretrained Transformers) have shown significant
advancements in natural language processing, offering more contextual understanding, we focused on LDA
for its established suitability in topic modeling. These alternative methods could be explored in future research
for a more detailed contextual analysis and to provide a comparison of their effectiveness in this domain (Barua
et al., 2014; Gurcan & Cagiltay, 2019).

LDA (Latent Dirichlet Allocation) is an unsupervised learning method that can be effectively applied to large
text collections without requiring a training set. Since this method does not require prior learning, large textual
datasets can be analyzed in a short time (Xu et al., 2022; Gurcan, 2023b). The LDA probabilistic topic
modeling algorithm involves a series of matrix operations performed on the DTM to reveal semantic
relationships between words (Egger & Yu, 2022). In this process, words with high co-occurrence rates are
grouped, with each word group representing a “topic”. The process of extracting semantic topics from
documents using the LDA approach is illustrated in Figure 1.

Figure 1. Extracting topics from documents using LDA

In LDA algorithms, Gibbs sampling is commonly used for parameter estimation. For the experimental analyses
in this study, the MALLET tool (McCallum, 2002), which implements a scalable version of the LDA model
based on Gibbs sampling, was used. MALLET, widely utilized in topic modeling applications, is a fast and
scalable Java-based tool that employs Gibbs sampling for topic modeling (Blei et al., 2003). In the analysis
performed using MALLET, the LDA algorithm was applied to the dataset with 1,000 iterations of Gibbs
sampling, resulting in the identification of 52 meaningful topics. The number of topics for the LDA model was
determined using coherence scores, which measure the semantic similarity of words within topics. Through an
iterative evaluation process, we identified 52 topics as the optimal number, balancing coherence and
interpretability.

200
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

3. Results and discussion

In this article, an experimental study was conducted using an LDA-based probabilistic topic modeling
approach to identify expert knowledge and skills related to emerging trends in software development. The
analysis was performed on a dataset derived from job postings for software developers. As a result of the topic
modeling analysis on the experimental dataset, 52 topics were identified, revealing the most significant
knowledge and skills associated with software development trends. These 52 topics, which represent trends in
software development expertise, were organized into a taxonomy related to core disciplines and areas of
software development expertise. Through this process, four main competency areas for expert knowledge and
skills in software development trends were identified. Subsequently, the identified 52 topics were assigned to
these areas based on their content, considering the keywords and percentage distributions within each topic.
The percentage distribution of topics across the four main competency areas is as follows: technical knowledge
and skills (23%), software development and programming skills (52%), analytical and logical skills (12%),
and non-technical skills (13%). Additionally, the distribution of topics across these competency areas in terms
of topic count is as follows: 13 topics for technical knowledge and skills, 22 topics for software development
and programming skills, 10 topics for analytical and logical skills, and 7 topics for non-technical skills.

At first glance, the findings indicate that the knowledge and skills required in the software development
industry encompass a broad range of disciplines. According to the results, software development expertise
demands not only comprehensive software development skills but also technical domain knowledge, analytical
and logical abilities, and individual (non-technical) skills such as communication. To provide a more detailed
understanding of the knowledge and skills associated with emerging software development trends, and to
further elaborate on the identified topics and competency areas, the findings are presented under four
subheadings.

3.1. Technical knowledge and skills

Among the 52 identified topics, 13 were observed to fall under the “technical knowledge and skills” category,
with a total frequency percentage of 23%. These 13 topics, assigned to the technical knowledge and skills area,
are presented in Figure 2 along with their assigned topic names and percentage values. The findings related to
technical knowledge and skills emphasize the theoretical knowledge areas that are essential for software
expertise.

The most prominent topic is “Educational requirements” (4.90%), which indicates the strong emphasis on
formal qualifications and foundational knowledge as key components in the competency framework.
Following this, “Database knowledge” (2.70%) and “Deep domain knowledge” (2.60%) are highly
emphasized, reflecting the importance of expertise in managing structured data and specialized knowledge in
specific domains. “Software architectures” (2.10%) underscores the significance of designing robust and
scalable systems, while “Software engineering knowledge” (1.80%) and “Software development processes”
(1.70%) highlight the need for understanding software lifecycle management and best practices.

Topics such as “Big data processing” (1.60%), “Scaling systems” (1.20%), and “Cloud-based systems”
(1.20%) represent emerging technical areas, focusing on the ability to handle large-scale data and scalable
infrastructures. These reflect the growing reliance on cloud technologies and distributed systems in modern
applications. Lesser-emphasized topics, including “Business intelligence” (0.90%), “Artificial intelligence”
(0.80%), “Data security and privacy” (0.80%), and “System administration” (0.70%), indicate specialized
areas of interest that, while not as dominant, are still essential for specific roles and applications in the field.

Overall, Figure 2 demonstrates a balanced distribution of foundational and advanced competencies, with a
clear focus on formal education, domain-specific expertise, and modern technical proficiencies. This
distribution reflects the diverse skill set required for professionals to succeed in the evolving landscape of
software engineering and technology. The results highlight that technical domain knowledge, such as database
expertise, software engineering principles, and software architectures, remains a fundamental requirement for
software professionals, emphasizing the theoretical foundation necessary for specialized roles in the industry.

201
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

Figure 2. Topics and rates for technical knowledge and skills

The focus on higher education, particularly in areas like computer science and engineering, suggests that
formal academic training continues to be highly valued, playing a crucial role in developing essential technical
knowledge for software development. The identified trends indicate that technical domain knowledge forms
the backbone of software expertise, with critical concepts such as software architectures and engineering
methodologies being consistently demanded by employers in the field.

3.2. Software development and programming skills

Of the 52 topics identified in the analysis, 22 were categorized under the broad competency area of “software
development and programming skills”, which carried a significant 52% weight across the overall distribution.
These 22 topics, detailed in Figure 3, are accompanied by their assigned topic names and corresponding
percentage values. This breakdown provides a clear understanding of the thematic focus within this
competency area, emphasizing its prominence in the software industry. The findings underline a critical
emphasis on application development, programming, and coding expertise, as well as the specific knowledge
areas required for success in the field of software development. Notably, new-generation scripting languages
and leading programming languages such as Java, Python, and JavaScript stand out as the most emphasized
qualifications. These findings highlight the importance of proficiency in versatile and widely used
technologies, reinforcing their relevance in modern software development environments.

The most prominent topics include “Scripting programming languages” (4.70%), “Javascript frameworks”
(4.40%), “Mobile development” (4.20%), “Application development” (4.10%), and “Web services
development” (3.80%). These topics represent the core technical competencies demanded by the industry.
“Scripting programming languages” and “Javascript frameworks” emphasize modern tools for front-end and
back-end development, while “Mobile development” and “Web services development” highlight the growing
need for scalable, cross-platform, and user-centric applications. Topics such as “Multi-platform development”

202
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

(3.40%), “Object-oriented programming” (3.20%), and “Java programming” (3.10%) reflect the importance
of foundational programming principles and creating solutions compatible with diverse platforms.

Figure 3. Topics and rates for software development and programming skills

Additionally, “Cloud-based development” (2.30%) and “Software design” (2.20%) emphasize the industry's
reliance on cloud architectures and structured software systems. Certain topics, such as “Real-time
programming” (2.10%) and “Open-source development” (1.90%), reflect specialized areas of expertise
valuable in specific domains. The inclusion of “Quality assurance testing” (1.80%), “Configuration control”

203
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

(1.70%), and “Agile development model” (1.50%) highlights the importance of testing, effective project
management, and iterative development processes. Emerging areas like “Big data development” (1.10%),
“Full-stack development” (1.10%), and “Troubleshooting” (1.00%) indicate growing but less dominant
demands for data-driven applications and end-to-end development. Similarly, “Scrum development model”
(0.90%), while less frequently mentioned, reflects its ongoing relevance as a widely adopted project
management methodology.

This distribution emphasizes a balanced focus on both core programming skills and modern development
approaches. The higher weight assigned to topics such as “Scripting programming languages”, “Javascript
frameworks”, and “Mobile development” aligns with current trends emphasizing application scalability, user
experience, and adaptability to emerging technologies. Additionally, the combination of foundational
programming principles (e.g., “Object-oriented programming”) with specialized areas (e.g., “Real-time
programming”) reflects the diverse skill set required in software development today. Overall, the topics
provides valuable insights into the technical competencies needed for professionals to remain competitive in
the dynamic software industry. The findings emphasize that proficiency in programming, paired with an in-
depth understanding of contemporary development methodologies, forms the backbone of technical expertise
in software development. The ability to utilize scripting languages effectively, develop mobile and web
applications, and implement modern programming frameworks is not only a testament to technical competence
but also a crucial determinant of professional success in today’s technology-driven landscape.

3.3. Analytical and logical skills

Among the 52 identified topics, ten were observed to fall under the category of “analytical and logical skills.”
The total percentage of these topics is 12%. These ten topics, assigned to the analytical and logical skills
competency area, are presented in Figure 4, along with their assigned topic names and percentage values. The
bar chart in Figure 4 presents the results of a topic analysis conducted on job postings, focusing on analytical
and logical skills. The Figure highlights “Requirements analysis” as the most commonly mentioned topic,
representing 2.20% of the analyzed postings. This suggests that employers place significant emphasis on
understanding and defining project or business requirements, which forms the foundation for effective
problem-solving and solution development.

“Data modeling and analysis” follows closely at 1.90%, underscoring the demand for professionals who can
structure and interpret complex data sets to inform business decisions. Similarly, “Analytical background”
(1.40%) ranks third, reflecting the high value placed on general analytical aptitude across industries. “Machine
learning methods” (1.30%), “Problem-solving skills” (1.20%), and “Software metrics” (1.10%) demonstrate
the growing importance of technical and quantitative competencies. These skills are particularly relevant in
roles that involve leveraging technology and data to optimize processes and outcomes.

On the other hand, “Predictive analytics” (1.00%), while less frequent, still highlights the need for forward-
looking insights in decision-making. Lower-ranked topics such as “Decision-making skills” (0.80%),
“Business analytics” (0.70%), and “Real-time analytics” (0.40%) suggest a more specialized focus, likely
relevant to specific roles or industries. This analysis reveals that job postings for analytical and logical skills
prioritize foundational technical abilities, such as requirements gathering and data handling, while also
emphasizing the importance of advanced techniques like machine learning and predictive analytics. However,
competencies related to real-time decision-making and business-oriented analytics appear less commonly,
potentially indicating niche demand or evolving trends in the job market.

These findings underscore the critical role of analytical and logical skills in shaping the modern workforce,
particularly in software development processes. Employers prioritize candidates who not only exhibit strong
logical reasoning and problem-solving capabilities but also possess expertise in leveraging data-driven tools
and methodologies to tackle complex challenges effectively. Key competencies such as requirements analysis
and data modeling emerge as foundational elements, reflecting their importance in addressing the multifaceted
demands of modern projects. Furthermore, the ability to adapt to evolving technologies and collaborate within
multidisciplinary teams has become increasingly essential. As industries continue to embrace digital
transformation, the demand for professionals who can combine technical expertise with strategic thinking is
likely to grow, further emphasizing the importance of these skills in shaping future career paths.

204
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

Figure 4. Topics and rates for analytical and logical skills

Additionally, skills like data visualization and various analytical methods are integral to enhancing software
development workflows, ensuring precision and efficiency in decision-making. The emphasis on advanced
techniques, including machine learning methods and predictive analytics, highlights the evolving nature of
expertise required in contemporary software development. This shift demonstrates the growing need for
professionals who can seamlessly integrate emerging technologies with traditional analytical competencies to
meet the dynamic demands of today’s industries.

3.4. Non-technical skills

It was discovered that seven of the 52 topics could possibly be classified as “non-technical skills”. These topics
make up 13% of the total. Figure 5 lists the seven topics in the non-technical skills competency category, along
with the topic names and percentage values. The findings related to non-technical individual skills emphasize
various personal competencies that are essential in software development roles, such as communication skills,
work experience, teamwork, critical thinking, and creativity.

Figure 5 illustrates the distribution of key interpersonal and non-technical skills identified in job postings. The
analysis reveals that “Communication skills” are the most emphasized, accounting for 5.10%, indicating their
paramount importance across industries. Effective communication is often critical for collaboration, client
interactions, and team coordination, making it a cornerstone of professional success. Following this, “Working
experience” ranks second at 2.90%, highlighting employers’ preference for candidates with practical, hands-
on experience. This suggests that prior experience is highly valued, likely reflecting the importance of applied
knowledge and familiarity with workplace dynamics. “Teamwork skills” (1.80%) emerge as the third most
mentioned topic, reinforcing the significance of collaborative abilities in achieving organizational goals. This
is complemented by “Critical thinking and creativity” (1.10%), which points to a growing demand for
innovative problem-solving and strategic thinking in the workplace.

Other notable mentions include “Project management” (0.80%), “Motivation and engagement” (0.70%), and
“Organizational skills” (0.60%), indicating that while these skills are less frequently highlighted, they remain
essential for certain roles, particularly in leadership and operational contexts. This analysis highlights a holistic
approach to skill requirements in job postings, specifically in the software development domain, where
interpersonal, creative, and professional competencies intersect to address the rapidly evolving demands of the

205
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

field. Among these, communication skills are identified as the backbone of effective collaboration in software
development. In this highly interdisciplinary domain, developers must articulate complex technical concepts
to both technical and non-technical stakeholders, ensuring that requirements are clearly understood and
solutions are aligned with organizational goals. Strong communication fosters teamwork, reduces
misunderstandings during critical development stages, and enhances project outcomes. Furthermore, in agile
and cross-functional development environments, the ability to adapt communication styles to diverse team
members and contexts is essential for success.

Figure 5. Topics and rates for non-technical skills

Equally critical in software development are critical thinking and creativity, which drive innovation and
adaptability. Critical thinking enables software professionals to evaluate problems systematically, analyze vast
amounts of data, and identify optimal solutions for highly complex and technical challenges. Whether
debugging code, optimizing algorithms, or designing system architectures, critical thinking ensures precision
and efficiency, helping teams deliver reliable and scalable software solutions. On the other hand, creativity
plays an essential role in fostering innovation within the development lifecycle. It allows developers to think
outside conventional frameworks, leading to the creation of groundbreaking software, novel user experiences,
and enhanced functionality. Creativity is not limited to front-end or user-facing design; it also drives innovation
in areas such as system optimization, automation, and backend efficiency. In an industry characterized by
continuous technological advancements, creativity empowers teams to remain competitive by devising unique
and forward-thinking solutions.

Together, communication, critical thinking, and creativity form the cornerstone of success in software
development. Their interconnected nature ensures that professionals can navigate the multifaceted challenges
of the field, effectively collaborate with multidisciplinary teams, and deliver innovative, user-centered
solutions. As software development becomes increasingly integral to the success of industries worldwide, these
skills are not just desirable—they are indispensable for driving progress, meeting complex project demands,
and advancing organizational objectives in this dynamic and competitive landscape.

4. Conclusions

In this study, the rapidly changing dynamics of the software industry and the evolving skill requirements
associated with it have been comprehensively addressed. Particularly in recent years, the rapid advancement
of technology has transformed the competencies required in software development processes, giving rise to a
multidisciplinary framework. Using the LDA-based probabilistic topic modeling method, analysis conducted
on a large dataset obtained from online job postings revealed that software developers increasingly require not
only technical skills but also soft skills such as analytical thinking, problem-solving, teamwork,
communication, and leadership. The findings emphasize the critical role of communication skills, which are
foundational in enabling effective collaboration across diverse teams, ensuring clarity in requirements analysis,
and facilitating productive exchanges between technical and non-technical stakeholders. Similarly, critical

206
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

thinking and creativity have emerged as indispensable for tackling the complex challenges of software
development, driving innovation, and optimizing solutions. These skills complement technical expertise by
empowering developers to analyze problems strategically and approach them with novel perspectives, leading
to better software solutions and advancements in system design. This comprehensive skill set, blending
technical, programming, analytical, and interpersonal competencies, is crucial for addressing the diverse and
complex challenges of the software development field. Both academic institutions and industry leaders should
recognize the importance of nurturing this diverse skill set to ensure that developers are well-prepared for the
future demands of the profession.

Although the findings of this study provide significant contributions, certain limitations exist. Firstly, the
dataset was obtained solely from online job postings, which may not fully reflect the general skill demands of
the entire software industry. Additionally, the LDA-based probabilistic topic modeling method may not
entirely capture the contextual depth of the topics while revealing latent structures in the texts, potentially
overlooking some essential skills. Finally, the dataset used in the analysis pertains to a specific time period,
making it challenging to fully capture the continuously evolving trends in software development. Therefore,
further studies conducted with more comprehensive and up-to-date data would enhance the validity of the
findings. Future research incorporating broader and more updated datasets will enhance the understanding of
skill requirements and support educational institutions, employers, and software developers in adapting to the
evolving demands of the industry. With further refinements, including expanding the dataset to incorporate
additional sources such as LinkedIn, Glassdoor, and Indeed, and enhancing the discussion of specific trends,
this work has the potential to become a leading contribution in its field. Additionally, the analysis could be
expanded by integrating deep learning-based NLP methods, such as BERT or GPT, to deepen the analysis and
enhance its granularity.

This study serves as a critical reference for understanding the current and future skill requirements in the
software industry. It offers valuable guidance to software developers in planning their careers by identifying
in-demand skills and knowledge areas. Employers can use these findings to better define workforce needs,
optimize employee selection processes, and design targeted training programs for skill development. For
educational institutions, the study provides actionable insights for updating curricula and developing training
programs that align with industry demands, ensuring that graduates are equipped with relevant competencies.
Additionally, the findings highlight emerging trends in software development, enabling individuals and
organizations to adapt to the evolving landscape of the industry. By addressing the unique needs of these key
stakeholders, the study acts as a strategic guide to foster growth, innovation, and alignment with next-
generation trends in software development.

Acknowledgement

This study is derived from the doctoral thesis by Fatih Gurcan titled “Determination of Expert Knowledge and
Skills for New Generation Software Development Trends Using a Probabilistic Topic Modeling Approach”,
conducted in the Department of Computer Engineering at the Institute of Science, Karadeniz Technical
University.

Author contribution

Fatih GURCAN – Responsible for conceptualization, methodology design, data collection, formal analysis,
writing, reviewing, and editing. Also, he contributed to data interpretation, statistical analysis, task
coordination, and validation of the findings. Cemal KOSE – Provided conceptual guidance and acted as an
advisor, contributed to software development, data visualization, curation, formal analysis, and assisted with
writing, reviewing, and editing.

Declaration of ethical code

The authors of this article declare that the materials and methods used in this study do not require ethics
committee approval and/or legal-special permission.

Conflicts of interest
The authors declare that there is no conflict of interest.

207
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

References

Aken, A., Litecky, C., Ahmad, A., & Nelson, J. (2010). Mining for computing jobs. IEEE Software, 27(1), 78–85.

Barua, A., Thomas, S. W., & Hassan, A. E. (2014). What are developers talking about? An analysis of topics and trends
in Stack Overflow. Empirical Software Engineering. [Link]

Blei, D. M. (2012). Probabilistic topic models. Communications of the ACM, 55(4), 77–84.

Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet allocation. Journal of Machine Learning Research, 3(4–
5), 993–1022. [Link]

Burkhardt, S., & Kramer, S. (2019). Decoupling sparsity and smoothness in the dirichlet variational autoencoder topic
model. Journal of Machine Learning Research, 20.

Chen, T. H., Thomas, S. W., & Hassan, A. E. (2016). A survey on the use of topic models when mining software
repositories. Empirical Software Engineering, 21(5), 1843–1919. [Link]

De Mauro, A., Greco, M., Grimaldi, M., & Ritala, P. (2018). Human resources for Big Data professions: A systematic
classification of job roles and required skill sets. Information Processing and Management, 54(5), 807–817.
[Link]

Debortoli, S., Müller, O., & Vom Brocke, J. (2014). Comparing business intelligence and big data skills: A text mining
study using job advertisements. Business and Information Systems Engineering, 6(5), 289–300.
[Link]

Egger, R., & Yu, J. (2022). A Topic Modeling Comparison Between LDA, NMF, Top2Vec, and BERTopic to Demystify
Twitter Posts. Frontiers in Sociology, 7. [Link]

Feng, J., Zhang, Z., Ding, C., Rao, Y., Xie, H., & Wang, F. L. (2022). Context reinforced neural topic modeling over
short texts. Information Sciences, 607, 79–91. [Link]

Gurcan, F. (2023a). Identification of mobile development issues using semantic topic modeling of Stack Overflow posts.
PeerJ Computer Science, 9, 1–28. [Link]

Gurcan, F. (2023b). What issues are data scientists talking about? Identification of current data science issues using
semantic content analysis of Q&A communities. PeerJ Computer Science, 9, e1361. [Link]
cs.1361

Gurcan, F., & Cagiltay, N. E. (2019). Big Data Software Engineering: Analysis of Knowledge Domains and Skill Sets
Using LDA-Based Topic Modeling. IEEE Access, 7, 82541–82552.
[Link]

Gurcan, F., & Kose, C. (2017). Analysis of software engineering industry needs and trends: Implications for education.
International Journal of Engineering Education, 33(4), 1361–1368.

McCallum, A. K. (2002). MALLET: A Machine Learning for Language Toolkit. In Http://[Link].

Moreno, A. M., Sanchez-Segura, M. I., Medina-Dominguez, F., & Carvajal, L. (2012). Balancing software engineering
education and industrial needs. Journal of Systems and Software, 85(7), 1607–1620.
[Link]

Stack Overflow. (2024). Stack Overflow Developer Jobs. [Link]

Terblanche, C., & Wongthongtham, P. (2015). Ontology-based employer demand management. Software: Practice and
Experience. [Link]

Thomas, S. W., Adams, B., Hassan, A. E., & Blostein, D. (2014). Studying software evolution using topic models. Science
of Computer Programming, 80(PART B), 457–479. [Link]

Vayansky, I., & Kumar, S. A. P. (2020). A review of topic modeling methods. Information Systems, 94.
[Link]

208
 Gürcan & Köse, 2025 • Volume 15 • Issue 1 • Page 197-209

Wu, X., Nguyen, T., & Luu, A. T. (2024). A survey on neural topic models: methods, applications, and challenges.
Artificial Intelligence Review, 57(2). [Link]

Xu, K., Lu, X., Li, Y. fang, Wu, T., Qi, G., Ye, N., Wang, D., & Zhou, Z. (2022). Neural Topic Modeling with Deep
Mutual Information Estimation. Big Data Research, 30. [Link]

209