Resume Parser Using Natural Language

Processing
Chris Jero Harsha P Rithik Harihar B S Dharun R
[Link] Computer Science [Link] Computer Science [Link] Computer Science
Engineering Engineering Engineering
SRM Institute of Science and SRM Institute of Science and SRM Institute of Science and
Technology, Technology, Technology,
Tiruchirappalli, India Tiruchirappalli, India Tiruchirappalli, India
chrisjeroharshax@[Link] bs.rithik07@[Link] dharundhoni28@[Link]
Abstract-In today's recruitment reducing the need for human
landscape, organizations receive involvement in shortlisting candidates.
thousands of resumes in different
formats. This makes manual screening Keywords-Resume Parsing, Natural
and data extraction slow and prone to Language Processing (NLP), Information
mistakes. This project proposes an Extraction, Python, MySQL, HTML,
automated Resume Parser that utilizes Named Entity Recognition (NER), Text
Natural Language Processing (NLP) to Mining, Data Visualization, Recruitment
extract key candidate information, Automation.
including name, contact details, I. INTRODUCTION
education, experience, and skills, from
unstructured documents. The system In today's rapidly changing digital world,
utilizes Python as its primary automation and data-driven intelligence
programming language, HTML for the are key to transforming traditional
web interface, and MySQL as the workflows. One major challenge in
backend database to store the parsed data recruitment is the manual processing of
efficiently. The model utilizes NLP resumes. This process involves identifying,
techniques, such as tokenization, part-of- reading, and sorting through unstructured
speech tagging, and named entity information from a large pool of
recognition, to convert the text in candidates. Manual screening is time-
resumes into a format that machines can consuming and can lead to human bias and
understand. The project also compares inconsistency. To address these issues,
the new NLP-based parser with Natural Language Processing (NLP) has
traditional keyword-based parsers, become a useful method for automating
showing improvements in precision and the extraction of structured information
efficiency. Experimental results are from unstructured text.
presented using pie charts and bar The Resume Parser using NLP project
graphs, highlighting significant gains in aims to create a web-based system that
data extraction accuracy. The proposed automatically gathers key details, such as
system improves the recruitment process the candidate's name, contact information,
by automating resume processing, educational background, skills, and
making it more scalable and precise experience from uploaded resumes. The
while system uses Python for programming,
HTML for the user interface, and MySQL
for managing the backend database. NLP
techniques like tokenization, part-of- Early resume parsing tools mostly used
speech tagging, and named entity rule-based or keyword-driven methods to
recognition (NER) assist in interpreting extract information with regular
text and converting it into a structured expressions and manual templates. These
format for storage and analysis. systems worked well with structured
resumes but had trouble with different
Unlike traditional keyword-based parsers
formats, inconsistent section names, and
that depend on strict patterns and fixed
varied language patterns. Early
templates, the proposed system adapts to
frameworks, like XML-based parsers and
different resume formats and various
regular-expression extractors, lacked
language structures. It improves accuracy
flexibility and needed frequent updates.
by employing contextual understanding
While they were simple to compute, they
and machine learning models from NLP
had low accuracy and high maintenance
libraries. To assess performance, the
costs, leading to a move toward NLP-
system compares its results with traditional
driven methods.
parsers and presents the findings using pie
charts and bar graphs, showcasing B. NLP-Based Information Extraction
improvements in accuracy, efficiency, and
With the growth of NLP and text mining
detection rates.
techniques, researchers started using
This project enhances intelligent tokenization, part-of-speech (POS)
recruitment systems by combining NLP- tagging, and named entity recognition
driven automation with web technologies (NER) to pull useful information from
to simplify resume analysis. The use of unstructured resumes. Tools like spaCy,
Python, HTML, and MySQL makes the NLTK, and Stanford CoreNLP became
system scalable and user-friendly, offering popular for creating effective parsers.
recruiters and organizations an efficient, Recent studies have incorporated deep
accurate, and real-time tool for evaluating learning models such as Bidirectional
candidates. LSTM, CRF (Conditional Random Fields),
and transformer-based structures like
II. RELATED WORK
BERT to identify complex entities such as
The automation of resume analysis has skills, job titles, and education details.
changed a lot because of improvements in While these models improved extraction
Natural Language Processing (NLP), accuracy, they required substantial
machine learning, and web technologies. computing power and labeled datasets.
Many studies and systems have been
C. Database-Driven Resume Management
developed to make parsing resumes in
Systems
different formats more accurate, flexible,
and efficient. This section looks at the key Data persistence and structured storage are
developments, methods, and frameworks essential parts of resume processing
that have influenced modern resume systems. Most of these systems rely on
parsers. It focuses on NLP-based relational databases like MySQL or
extraction techniques, data storage PostgreSQL to keep organized records of
methods, and ways to compare models. extracted resume data. Some modern
frameworks use NoSQL databases like
MongoDB for flexible storage of semi-
A. Traditional Resume Parsing Systems structured text. Using databases enhances
retrieval efficiency. This setup allows illustrating the results through graphs, the
recruiters to conduct filtered searches, system shows clear improvements in
build candidate profiles, and link resumes accuracy and usability.
to job requirements. However, scalability
III. METHODOLOGY
and query optimization continue to be
significant challenges, particularly with The Resume Parser project is built on a
large datasets. modular system that uses NLP. It aims for
accuracy, scalability, and ease of use.
D. Comparative and Hybrid Parsing
Development went through five stages:
Techniques
requirement analysis, system design,
Several recent studies have explored implementation, testing, and final
hybrid methods that combine NLP with evaluation. This structured approach
rule-based validation to boost reliability. helped ensure the web-based parser is
These systems use machine learning reliable and efficient.
models for extracting entities and apply
During the requirement analysis stage, we
heuristic rules for validation. This
aimed to automate the extraction of useful
approach increases accuracy and reduces
information from resumes using Natural
false positives. Comparative research
Language Processing. The functional
between traditional parsers and NLP-based
requirements included creating a web
models shows that NLP methods perform
interface for users to upload resumes,
better than static models when handling
building a Python-based backend to
various resume formats and unstructured
process and extract resume data, and using
text. Some studies also include statistical
a MySQL database to organize the
evaluation metrics like precision, recall,
extracted details. We also looked at non-
and F1-score to evaluate extraction quality,
functional requirements, focusing on good
forming the basis for this paper’s
system performance, data protection, and
comparative analysis.
the ability to handle larger or repeated use.
E. Summary of Literature Insights
In the system design and implementation
The literature indicates that resume parsing stages, we developed the frontend with
has progressed from simple rule-based HTML and CSS to create a clean, user-
systems to smart, data-driven, and flexible friendly interface for uploading resumes
frameworks. Despite the advancements, and viewing results. The backend, made in
challenges remain in managing multi- Python, performs key NLP tasks like
format documents, unclear section headers, tokenization and Named Entity
and multilingual content. Combining Recognition (NER). We sorted the
natural language processing with solid extracted information into main categories
database systems and user-friendly web such as Name, Email, Education, Skills,
interfaces offers a promising way for and Work Experience. This organized data
scalable and effective resume analysis. The is then stored in a MySQL database for
proposed system adopts this approach by future display or analysis. The complete
utilizing Python-based NLP techniques, workflow of the system is shown in Fig. 1.
MySQL storage, and an HTML frontend to
deliver a complete intelligent resume
parsing solution. Also, by comparing its
performance to a traditional parser and
 not only minimizes human effort but also
supports faster and more informed
decision-making in recruitment
workflows..
IV. IMPLEMENTATION
The Resume Parser system was built as a
full-stack web application using Python,
Flask, HTML, and a backend database.
This setup transforms the design concept
into a functioning model that automatically
extracts, processes, and organizes resume
information. The backend runs on Python
and Flask, which manage tasks like
handling file uploads, routing user
Fig. 1. Workflow diagram of the proposed
requests, and returning processed results.
Resume Parser system.
The NLP component relies on libraries
The system was developed and tested such as spaCy, NLTK, and regular
using Visual Studio Code (VS Code) as expressions to perform tokenization and
the primary development environment, Named Entity Recognition (NER). This
which streamlined coding, debugging, and allows the system to identify key details
project organization. To verify its like Name, Contact Information,
robustness, the team conducted both Education, Skills, and Work Experience
functional and accuracy testing using from various resume formats.
resumes of different formats, layouts, and
On the frontend, HTML and CSS create a
levels of detail. This ensured that the
clean, user-friendly interface for users to
parser could handle real-world variations
upload resumes and view the extracted
rather than performing well only on
data instantly. When a resume is
uniform templates.
submitted, Flask processes the file in real
A comparative analysis was then carried time and displays the structured
out between the proposed NLP-based information directly on the interface. The
parser and a traditional keyword-based extracted data is stored in an organized
parser. The results clearly showed that the database structure that allows for smooth
NLP-driven system delivered more retrieval and updates whenever needed.
accurate data extraction and significantly
Overall, this implementation shows
reduced the need for manual review or
smooth communication between the web
correction. To better illustrate these
interface, the NLP processing layer, and
findings, the performance metrics were
the data storage component. It ensures that
visualized using pie charts and graphs,
resume information is extracted correctly
which made the improvements easier to
and presented in a clear, structured format
interpret and evaluate.
without needing manual review. This
Overall, the evaluation demonstrated that foundation allows for deeper analysis and
the system successfully automates the visualization of the extracted data, which
resume screening process with improved will be explored in the following section.
reliability, efficiency, and consistency. It
 V. RESULT AND DISCUSSION VI. CONCLUSION
The Resume Parser uses NLP technology The Resume Parser using NLP system
to automatically extract candidate details automates the extraction and organization
from resumes through a web platform. of information from resumes. This cuts
Users can upload resumes in different down on the need for manual screening
formats using the HTML interface. The during recruitment. By using Python NLP
Flask backend and Python NLP module techniques, Flask, HTML, and MySQL,
handle the processing of these resumes. the system ensures a smooth data flow
The system identifies key information such from upload to storage and visualization.
as Name, Contact, Education, Skills, and NLP methods like tokenization and named
Experience, which is stored in a MySQL entity recognition help clearly identify
database for easy management and important fields, including Name, Skills,
retrieval. Education, and Experience across different
resume formats.
We tested multiple resume formats to
ensure accuracy and consistency. The Unlike traditional keyword-based parsers,
parser displayed great flexibility with this model provides better accuracy and
various layouts and text styles, accurately flexibility in handling unstructured text. It
extracting important information across also includes a user-friendly web interface
different templates. The results are shown and secure database storage, making it
in a clear format on the web page for users ideal for real-time recruitment and HR
to review and check. applications. Future improvements could
involve machine learning-based ranking
and support for multilingual resumes to
improve its usability and reach.
REFERENCES

[1] A. Kumar and S. Gupta, “Automated

Resume Screening Using Natural
Language Processing,” International
Journal of Computer Applications, vol.
183, no. 42, pp. 25–29, 2022.

[2] P. Sharma, R. Jain, and M. S. Bansal,

Fig. 2. Sample output showing extracted “Information Extraction from Resumes
resume information displayed on the web Using Named Entity Recognition,” IEEE
interface. International Conference on
Computational Intelligence and
We compared the NLP-based parser with a Communication Networks (CICN), pp.
traditional keyword-based parser. The 650–654, 2021.
NLP model achieved higher precision in
finding relevant entities and required less [3] M. Arora and D. Singh, “Design and
manual correction. Overall, the system Development of Resume Parser using
proved efficient, reliable, and suitable for Python NLP and Machine Learning,”
automated recruitment processes. International Journal of Advanced
Research in Computer Science (IJARCS),
vol. 12, no. 6, pp. 45–50, 2021.
[4] S. Chatterjee and P. Ghosh,
“Application of Natural Language
Processing in Recruitment Automation,”
IEEE Access, vol. 10, pp. 52530–52539,
2022.

[5] T. Joshi, A. Patel, and K. Mehta, “Text

Mining and Information Extraction for HR
Automation,” Journal of Emerging
Technologies and Innovative Research
(JETIR), vol. 9, no. 2, pp. 310–315, 2023.

[6] Y. Li and J. Zhao, “A Deep Learning

Approach for Resume Information
Extraction,” Proceedings of the 2023
IEEE 17th International Conference on
Semantic Computing (ICSC), pp. 412–418,
2023.