Topic Description

Recommendations are ubiquitous in our daily life: Whenever we are visiting Amazon, Netflix or Google, a plethora of products, services, and movies are presented to us, to increase the time and money spent on the respective platforms. This trend has been fueled by the ever-growing amount of data available digitally; allowing the highly digitalized players to crunch the collected data about personal preferences, user characteristics and products into what is commonly coined “content-based” or “collaborative” recommendations [1]. Challenges like the prominent one “million dollar prize” advertised by Netflix for the person improving their recommendations the most [2] are indicators for the massive importance of such algorithms for large retailers and service providers.

While online business and e-commerce have been thriving on the technological advances of the last two decades, high streets and brick-and-mortar retailing faced a steady decline [3], [4]. To avoid inner-city wastelands, practitioners and academics set out to “update” the high street. One of these projects is smartmarket², where we seek to develop location-based services and apps for an interactive shopping experience in urban areas. After an initial field test of our smartmarket² platform in the city of Paderborn, we have data of 66 local business, and trajectory data of 400 participants collected with the help of more than 120 Bluetooth beacons across the high street of Paderborn. In effect, we have digital “customer journeys”, which are time-logical sequences of businesses passed-by and visited in the users’ high street visits. This data already helped us and local business owners to better understand how people move through Paderborn. The next step will be to offer app users recommendations based on the trajectories of likeminded peers and available information on the businesses. During the course of your thesis, you will conceptualize and implement a geo-recommender system based on the existing data set.

You have degrees of freedom in the choice of your technology stack for the geo-recommender system. Users will receive geo-recommendations during their shopping trips through the smartmarket²-app. Therefore, you will extend the existing smartmarket²-app for Apple iOS (no worries: We can provide you with the required hardware).

If you want to tackle a very recent and important topic at the confluence of service science, data analytics, and information retrieval, feel free to contact me to learn more about the topics and to discuss your next steps!

PS:       You should not be afraid of the implementation aspect – the exact extent and setup can be tailored to fit your interest and competences.

PPS:    Some preliminary work will be published on the ECIS 2019 in Stockholm - the preprint can be downloaded here.

Literature

[1]      G. Adomavicius and A. Tuzhilin, “Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions,” IEEE Trans. Knowl. Data Eng., vol. 17, no. 6, pp. 734–749, Jun. 2005.

[2]      J. Bennett and S. Lanning, “The Netflix Prize,” in Proceedings of the KDD-Cup and Workshop at the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2007.

[3]      L. Bollweg, R. Lackes, M. Siepermann, and P. Weber, “Carrot-or-Stick: How to Trigger the Digitalization of Local Owner Operated Retail Outlets?,” in Proceedings of the 51st Hawaii International Conference on System Sciences, 2018.

[4]      C. Bach, “IFH: Jedes zehnte Geschäft in Deutschland ist von der Schließung bedroht,” Location Insider, 2015. [Online]. Available: http://locationinsider.de/ifh-jedes-zehnte-geschaeft-in-deutschland-ist-von-der-schliessung-bedroht/. [Accessed: 09-Feb-2018].