Design and Development of a Self-Service BI&A Guided Recommendation System Based on Logfiles of the Bosch’s Corporate Data Lake

  • Problem Description

    The provisioning of the right information to the right person at the right time is critical to remain competitive and represents a key concern of Business Intelligence & Analytics (short, BI&A, Bucher et al., 2009). To ensure continued success, companies have widely implemented and integrate BI&A systems (Moore 2016). Consequently, more and more analytics are expected to be done by business users, but business users have significant difficulties to conduct for. e.g. ad-hoc analyses with existing BI&A systems (Mertens, 2013). Currently, the complexity can often be only handled by experienced (power) users which are therefore in high demand and become more and more a bottleneck for data-driven decision making. One branch of BI&A technologies addressing this challenge is called Self-Service BI&A. A major idea of SSBIA is that casual users will be able to prepare and analyze data with easy-to-use BIA systems without the need for expert support (Lennerholt et al., 2018). To increase the adoption of such systems, one recent research stream describes to capture and transfer the knowledge of power user by utilizing recommender system techniques like for example collaborative filtering (e.g see references). A data source could be log files from existing BI&A systems. With such a system, (not only) business users could be encouraged to use (Self-Service) BI&A system on their own, and with given recommendations, guided through the analysis process, named data science lifecycle (e.g. CRISP-DM).
    The thesis is conducted in close collaboration with the Robert Bosch GmbH. A contract with Bosch can not be offered. 

    Goal of Thesis

    The overall goal of the thesis is to establish a design science research project to design, build and evaluate a Self-Service BI&A guided recommendation system by considering the current state of the art in literature and available data. For the development of the system, SQL log files from the corporate data lake of the Robert Bosch GmbH are provided. Given recommendations shall guide users throught the data science lifecycle. The student can rely on existing guidance design features (e.g., Morana et al. 2017).

    Work Packages

    1. Conduct a systematic literature review to examine the state of the art of (technical) concept (for e.g. in the form of recommender systems) to encourage and guide users through the stages of the data science lifecycle
    2. Map existing concepts to the stages of the data science lifecycle
    3. Collect ideas for a guidance services by analyzing the logfiles provided
    4. Design and implement a Self-Service BI&A guided recommendation system.
    5. Evaluate your implementation. Potentially, a focus group discussion with Bosch experts can be arranged.

    Skills required

    • High intrinsic motivation and proper time management
    • Good programming skills in a common data science language such as Python, R or others as well as profound SQL knowledge
    • Some experience with recommendation systems  (e.g. from the lecture Recommender Systems @IISM)
    • Experience in frontend programming with JS and preferably Vue.js or other frameworks
    • Fluent English (as the thesis has to be written in English)

    Contact

    If you are interested, drop me an email with a short motivation statement, your CV and your current transcript of records. If you have questions before, do not hesitate to contact me.
    sven.michalczyk@kit.edu

    References

    Bucher, T., Gericke, A., and Sigg, S. (2009). “Process-centric business intelligence,” Business Process Management Journal (15:3), pp. 408-429.
    Lennerholt, C., Van Laere,J., and Söderström, E. (2018). “Implementation Challenges of Self Service Business Intelligence: A Literature Review,” In: Proceedings of the 51st Hawaii International Conference on System Sciences (HICSS), pp. 5055–5063.
    Mertens, M. 2013. “KNOBI - knowledge-based business intelligence for business user information-self-service”, OlWIR, Oldenburger Verl. für Wirtschaft, Informatik und Recht, Edewecht (2013)
    Moore, S. 2016. Gartner Says Worldwide Business Intelligence and Analytics Market to Reach $16.9 Billion in 2016, http://www.gartner.com/newsroom/id/3198917
    Morana, S., Schacht, S., Scherp, A., & Maedche, A. (2017). A Review of the Nature and Effects of Guidance Design Features. Decision Support Systems, 97, 31–42. https://doi.org/10.1016/j.dss.2017.03.003
    Examples of BI&A Recommendation Systems
    Dabab, M.; et.al (2018):  A Decision Model for Data Mining Techniques, Portland International Conference on Management of Engineering and Technology (PICMET), Honolulu, HI, 2018, pp. 1-8
    Drushku, Krista; Aligon, Julien; Labroche, Nicolas; Marcel, Patrick; Peralta, Verónika (2019): Interest-based recommendations for business intelligence users. In Information Systems 86, pp. 79–93.
    Lismont, J., Van Calster, T., and Óskarsdóttir, M. (2019). “Closing the Gap Between Experts and Novices Using Analytics-as-a-Service: An Experimental Study,” Business Information Systems Engineering, (6), pp. 1-15. (DOI: 10.1007/s12599-018-0539-z).
    Nargesian, F., Biem, A., Jain, P., Parthasarathy, S., & Turaga, D.S. (2015). SOFIA: An Analytics Recommendation System. International Semantic Web Conference.
    Niu, Li; Lu, Jie; Zhang, Guangquan; Wu, Dianshuang (2013): FACETS: A cognitive business intelligence system. In Information Systems 38 (6), pp. 835–862. DOI: 10.1016/j.is.2013.02.002.
    Sulaiman, S.; Gómez, J. M. (2018): Recommendation-based Business Intelligence Architecture to Empower Self Service Business Users. In : Multikonferenz Wirtschaftsinformatik.
    Zschech, Patrick et. al. (2019): Towards a Text-based Recommender System for Data Mining Method Selection. In : 25th Americas Conference on Information System. (AMCIS). Cancun.