This presentation will include the results of Big Data using social media data to predict impact on the university brand and resource allocations for achieving career-relevant programs for working adults. The results connect the university brand attribute of career-relevant programs to student perceptions and align with Knowles theory of andragogy.
Extended Abstract
Executive leaders in the highly competitive education industry are facing multiple challenges, which require new and innovative approaches to strategic planning and decision making. The rise of Big Data and Machine Learning practices provide unique opportunity for leaders in the higher education industry (HEI) to gain insights from data and create value for students, alumni, and faculty. The benefits include soaring with scientific revolution in lighting speeds for data collection and analysis that is cost-effective, secure, and provides efficient IT storage and processing capabilities via cloud distributed computing.
This study presents an innovative approach by applying Big Data with Machine Learning practices to collect and analyze social media data for impact to the university brand and resource allocations via the Amazon Web Services (AWS) platform. The study applies a quantitative approach to examine the quality of evidence using social media data from an academic social networking site and a university’s LinkedIn site to construct variables – Customer Experience (CX), Advocacy, and Persona – and to build a binary regression model to predict impact to brand and resource allocations in the context of achieving the university’s vision: recognition as a trusted provider of career-relevant programs for working adults.
It is anticipated this study will add new knowledge in research design methods for what is now possible with Big Data/Machine Learning practices and help inform internal processes on impact to the university brand and resource allocations. In addition, the exploratory nature of this study along with its innovative approach aligns with the need for dynamic leadership approaches to learn and to discover new insights from data for sustainable value creation and competitive advantage in the HEI.
From a business perspective, this study should be of interest to leaders in the HEI for purposes of improving decision making to create value for students, faculty, and alumni. It should be of particular interest for purposes of gaining new insights from Big Data/Machine Learning practices in the context of growing the university brand for career-relevant programs and employment.
Participants attending this research session will learn the following:
- Awareness in the practical application of Big Data using the AWS platform for higher education
- New research design methods for collecting, categorizing, and analyzing unstructured datasets
- The study’s results for predicting impact to university brand and resource allocations
Audience engagement will include the following techniques:
- Interactive handouts
- Interactive questions and answers
Materials to be provided during the presentation include:
- Power Point presentation
- Handouts
- Web Links
Who Might Benefit from the Presentation?
- Institution Type: Higher Education
- Audience Level: All
Problem Statement
At the present time, very little evidence exists in the literature on HEI use of Big Data with Machine Learning practices as applied to the impact on brand and resource allocations to analyze large data sets of social-media messages and to uncover trends for improving the Customer Experience (CX). This study will examine the sentiment of Customer Experience at a university to assess its brand in the context of career-relevant programs using social media data from the university’s proprietary social-networking sites. The objective is to build a predictive model and identify variables that could shift perceived value in a data point (favorable/unfavorable) with impact to the brand and resource allocations. The study will be completed in August, 2016.
Research Questions
There are two research questions to guide the researchers’ inquiry, each having corresponding hypotheses:
Q1: To what extent does social-media data from students, faculty, and alumni in XXXX database and LinkedIn predict the university brand attribute of being a trusted provider of career-relevant programs for working adults?
Q2: To what extent does social-media data from students, faculty, and alumni in XXXX database and LinkedIn predict impact upon university resource allocations to (1) buy/build assets or (2) sell/remove assets?
Hypotheses
H1 Customer Experience (CX) and willingness to advocate the university to others is moderated
by Persona.
H2 There is a relationship from the combined effects of Customer Experience (CX) and Advocacy to the
university brand attribute (career-relevant programs).
H3 There is a relationship from the combined effects of Customer Experience (CX) and Advocacy to
impact university resource allocations.
Figure 1: Conceptual Model of Big Data Use Case on University Brand Impact & Resource Allocation
Methods
This study is a quantitative case study. Quantitative analysis will be performed using advanced computing technologies to collect and categorize unstructured datasets and then perform binary classification of positive or negative messages for logistic (logit) regression.
Population
The student population of the university is XXXXX and it is growing with degrees and programs leading to employment. The university employs over XXXX full-time faculty and over XXX part-time faculty employees. Alumni members are over XXXXX. (Exact numbers will be provided in final presentation.)
Results
The details of data analysis results, discussion/interpretations, and conclusion are scheduled to be completed by August 31, 2016.
Conclusion
Big Data with Machine Learning practices represents the next evolution in leadership development to understand IT capabilities and be able to engage in meaningful conversations with data science team members (Provost & Fawcett, 2013) – that is, asking the right questions, and knowing what algorithms, from which machines learn, need to be tweaked for improved accuracy.
References (Partial List)
Baron, R.M., and Kenny, D.A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology 51(6), 1173-1182.
Bean, J., and Van Tyne S. (2012). The Customer Experience Revolution. Vermont: Bridgantine Media.
Chapleo, C., and Reader, P. (2014). Higher education brands and data. In Using Data to Improve Higher
Education (p. 81-91). Rotterdam/Boston/Taipei: Sense Publishers.
Knowles, M. (1984). Andragogy in Action. San Francisco: Jossey-Bass.
Lilien, G.L., Rangaswamy, A., Van Bruggen, G.H., and Starke, K. (2004).
DSS Effectiveness in marketing resource allocation decisions: Reality vs. perception.
Information Systems Research 15(3), 216-235.
Picciano, A.G., (2013). The evolution of big data and learning analytics in American higher education.
Journal of Asynchronous Learning Networks, 16(3), 9-20.
Provost, F. and Fawcett, T. (2013). Data Science For Business: What you need to know about
data mining and data-analytic thinking. O’Reilly: Sebastopol, CA.
Pyle, D., and Jose, C.S. (June 2015). An executive’s guide to machine learning. McKinsey Quarterly.
Retrieved at http://www.mckinsey.com/insights/high_tech_telecoms_internet/an_executives_guide_to_machine_learning
Romdhane, L.B., Fadhel, N., and Ayeb, B. (2009). Building customer models from business data: An
automatic approach based on fuzzy clustering and machine learning. International Journal of
Computational Intelligence and Applications 8(4), 445-465.
Sagar, S.D., Dehuri, S., and Wang, G. (2012). Machine learning for social network analysis: A systematic
literature review. The IUP Journal of Information Technology 8(4), 30-51.
Terkla, D. G., Sharkness, J., Conoscenti, L.M., and Butler, C. (2014). Using data to inform institutional
decision making at Tufts University. In Using Data to Improve Higher Education (p. 39- 63).
Rotterdam/Boston/Taipei: Sense Publishers
White, T. (2015). Hadoop: The definitive guide; storage and analysis at Internet scale.
O’Reilly: Sebastopal, CA.
Winship, C. and Mare, R.D. (August, 1984). Regression models with ordinal variables.
American Sociological Review 49, 512-525.