Exploiting Time Series Analysis in Twitter to Measure a Campaign Process Performance
Source of Publication
Proceedings - 2017 IEEE 14th International Conference on Services Computing, SCC 2017
© 2017 IEEE. While there are several metrics to measure business process performance, recently there is an additional requirement from businesses to evaluate business processes based on their impact on users. In this work, we evaluate business process performance using social media analytics. We view a marketing campaign as a business process and we evaluate its performance based on its impact on the Twitter. We propose a new way to calculate the 'follow' relationship in Twitter based on the users' reaction to the marketing campaign process activities and we use time series and sentiment analysis for defining and measuring performance. We re-build the Twitter graph based on users' reactions to the marketing activities in time and we are using community detection algorithms to identify the size of the 'follow' community and thus we define metrics to calculate the impact of the marketing/campaign process. We evaluate our approach using a dataset for a given politician. We re-construct the campaign process as a set of activities on specific topics (promotions) in time using LDA. Our results show that social media analytics can be used as a valid metric for assessing business processes performance.
Institute of Electrical and Electronics Engineers Inc.
Analytics, Community detection, Marketing campaign business process, Time series analysis
Kafeza, Eleanna; Makris, Christos; and Rompolas, Gerasimos, "Exploiting Time Series Analysis in Twitter to Measure a Campaign Process Performance" (2017). All Works. 1592.
Indexed in Scopus