Design and Implementation of an Integrated Analytics and Recommendation Framework for the Punjab Government Rozgar and Karobar Mission (PGRKAM) Platform

Abstract

The Punjab Government Rozgar and Karobar Mis-sion (PGRKAM) was created as a large-scale digital platform to connect job seekers and employers. While it has managed to attract a broad user base, its current format remains closer to a digital notice board than an intelligent job portal. This paper proposes a technical framework to improve PGRKAM by introducing advanced analytics and a hybrid recommendation system. The recommendation system makes use of Genetic Algorithms (GA) to match skills and qualifications with job requirements and Collaborative Filtering (CF) to learn from user behaviour patterns. To help administrators, the design also includes interactive dashboards that display trends in user activities, employer participation, and job market needs. The overall architecture is built to be scalable with data protection in mind, incorporating strong encryption to remain aligned with India’s Personal Data Protection Bill (PDPB). With these improvements, PGRKAM can move from being a static listing of jobs to a responsive and intelligent job marketplace, improving job placements and greatly enhancing user satisfaction.

Introduction

This paper proposes an AI-enhanced redesign of the Punjab Government's Rozgar and Karobar Mission (PGRKAM) platform by integrating an intelligent job recommendation engine and analytics dashboard. While the existing PGRKAM platform primarily functions as a job listing portal, it lacks advanced analytics, personalized recommendations, and employment intelligence, limiting its effectiveness for both job seekers and administrators. In contrast, private platforms such as LinkedIn and Indeed use machine learning to provide personalized job recommendations and user insights.

The proposed framework transforms PGRKAM into a state-level employment intelligence platform by combining Collaborative Filtering (CF) and Genetic Algorithms (GA) for intelligent job matching. Collaborative Filtering recommends jobs based on similarities in user behavior, while the Genetic Algorithm optimizes candidate-job matching using weighted criteria such as skills, experience, and location. This hybrid approach improves recommendation accuracy, personalization, diversity, and recall compared to using GA alone.

The architecture consists of two parallel pipelines:

• Analytics Pipeline: Captures user interactions (page views, sessions, demographics, traffic sources, etc.) through an analytics SDK similar to Google Analytics and provides behavioral insights.
• Recommendation Pipeline: Processes user profiles, job postings, and application history using the hybrid recommendation engine to generate personalized job suggestions and evaluate application outcomes.

Both pipelines feed into a unified Administrator Analytics Dashboard, enabling government officials to monitor:

• User acquisition and demographics.
• User engagement and application funnels.
• Performance of recommendation algorithms.
• Job market trends, including in-demand skills and occupations.
• Application success rates and employment statistics.

The proposed system uses a structured database containing:

• User profiles.
• Job listings.
• User interaction events.
• Job application history.

The implementation employs:

• Python with Scikit-learn and NumPy for the recommendation engine.
• PostgreSQL for data storage.
• Redis for caching.
• React or Angular with D3.js/Chart.js for interactive dashboard visualization.
• Integration with Google Analytics APIs for user behavior analytics.

Performance evaluation is based on Precision and Recall, measuring the accuracy and completeness of job recommendations. Comparative analysis shows that the hybrid GA + CF approach provides:

• Higher precision and recall.
• Better personalization by combining profile and behavioral information.
• Greater recommendation diversity and serendipity than GA alone.

Compared with standard analytics platforms such as Google Analytics, the proposed system additionally supports:

• AI-based job recommendations.
• Application success tracking.
• Government policy reporting.
• Localized employment analytics for Punjab.

The paper concludes that integrating intelligent recommendations with comprehensive analytics can transform PGRKAM from a basic job portal into a data-driven employment ecosystem that benefits job seekers, recruiters, and policymakers. Future work includes incorporating Natural Language Processing (NLP) to better analyze resumes and job descriptions and developing predictive models to forecast labor market trends and support evidence-based government policymaking.

Conclusion

In this paper, we’ve laid out a plan to give the PGRKAM platform a major upgrade by adding smart analytics and a custom-built AI for recommending jobs. The idea is to watch how people use the site and use that information to make much better job matches, giving the people in charge the insights they need to improve everything. Our main goal is to turn the site from a simple job board into a dynamic hub that genuinely helps the people of Punjab find great opportunities. Looking ahead, we want to make this system even smarter. We plan to use Natural Language Processing (NLP) to really dig in and understand the fine details in resumes and job postings. We also aim to build models that can predict future job market trends, which will give the government solid data to make proactive and well-informed policy decisions.

References

[1] M. Mujeerulla, Preethi, M. S. Khan, et al., “Demerits of elliptic curve cryptosystem with Bitcoin curves using Lenstra–Lenstra–Lovasz (LLL) lattice basis reduction,” Arab J. Sci. Eng., vol. 49, pp. 4109–4124, 2024. doi: 10.1007/s13369-023-08116-w [2] M. S. Khan, T. M. Chen, M. Sathiyanarayanan, M. Mujeerulla, and S. P. Raja, “Application of Lenstra–Lenstra–Lovasz on elliptic curve cryptosystem using IoT sensor nodes,” Journal of ICT Standardization, vol. 12, no. 4, pp. 381–408, 2025. doi: 10.13052/jicts2245-800X.1242 [3] Preethi, M. M. Ulla, S. R., and R. M. Devadas, “Blockchain modeled swarm optimized Lyapunov smart contract deep reinforced secure tasks offloading in smart home,” MethodsX, vol. 14, 2025, Art. no. 103305. doi: 10.1016/j.mex.2025.103305 [4] Preethi, M. M. Ulla, G. P. K. Yadav, K. S. Roy, R. A. Hazarika, and K. S. K., “Elliptic-Curve Cryptography Implementation on RISC-V Processors for Internet of Things Applications,” Journal of Engineering, vol. 2024,Art. ID 5116219, 2024. doi: 10.1155/2024/5116219 [5] M. M. Ulla, Preethi, M. S. Khan, S. L., and S. B. J., “Lightweight Strobe Security Libdisco Scheme for IoT-Based Sensor Data,” in Proc. 2024 8th Int. Conf. on CSITSS, 2024. doi: 10.1109/CSITSS64042.2024.10817029 [6] F. Ricci, L. Rokach, and B. Shapira, Recommender Systems Handbook. Springer, 2015. [7] Y. Koren, R. Bell, and C. Volinsky, “Matrix factorization techniques for recommender systems,” Computer, vol. 42, no. 8, pp. 30–37, 2009. [8] R. Burke, “Hybrid recommender systems: Survey and experiments,” User Modeling and User-Adapted Interaction, vol. 12, no. 4, pp. 331–370, 2002. [9] J. H. Holland, Adaptation in Natural and Artificial Systems. MIT Press, 1992. [10] D. Jannach, M. Zanker, A. Felfernig, and G. Friedrich, Recommender Systems: An Introduction. Cambridge Univ. Press, 2010. [11] S. Singh and S. Chatterjee, “E-governance initiatives in India: The case of National Career Service (NCS),” Gov. Info. Quarterly, vol. 36, no. 3, pp. 449–458, 2019. [12] S. Bhatnagar, E-Government: From Vision to Implementation. SAGE, 2004. [13] Google, “Google Analytics Documentation,” 2023. [Online]. Available: https://developers.google.com/analytics [14] Mixpanel, “Product Analytics Platform,” 2023. [Online]. Available: https://mixpanel.com [15] Kissmetrics, “Behavioral Analytics,” 2023. [Online]. Available: https://www.kissmetrics.io [16] K. B. C. Saxena, “Towards excellence in e-governance,” Int. J. Public Sector Management, vol. 18, no. 6, pp. 498–513, 2005.

Copyright

Copyright © 2026 Mohanavel M. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.