Exploring Insurance and Natural Disaster Tweets Using Text Analytics

Exploring Insurance and Natural Disaster Tweets Using Text Analytics

Tylor Huizinga (Brock University, St. Catharines, Canada), Anteneh Ayanso (Brock University, St. Catharines, Canada), Miranda Smoor (Brock University, St. Catharines, Canada) and Ted Wronski (Brock University, St. Catharines, Canada)
Copyright: © 2017 |Pages: 17
DOI: 10.4018/IJBAN.2017010101
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

This study explores twitter data about insurance and natural disasters to gain business insights using text analytics. The program R was used to obtain tweets that included the word ‘insurance' in combination with other natural disaster words (e.g., snow, ice, flood, etc.). Tweets related to six top Canadian insurance companies as well as the top five insurance companies from the rest of the world, including the new entrant Google Insurance, was collected for this study. A total of 11,495 natural disaster tweets and 19,318 insurance company tweets were analyzed using association rule mining. The authors' analysis identified several strong rules that have implications for insurance products and services. These findings show the potential text mining applications offer for insurance companies in designing their products and services.
Article Preview

Introduction

Advances in social media technologies have brought a new level of communication and interaction among Internet users (Anderson, 2007). However, much of the data on the Web today is unstructured that requires the use of advanced big data technologies in order to generate actionable insights (Abdelhafez, 2014; Bihl, Young II, & Weckman, 2016). As a powerful platform for self-publishing and interaction, social media has provided businesses with tremendous opportunity to establish a sense of participation from viewers and generate business insights (Gupta & Gupta, 2016). Fan and Gordon (2014) stated that 91% of adults online are using social media on a regular basis. As one of the most widely used platforms today, Twitter allows users to express their thoughts, emotions and facts within a 140 character base to other users who are interested and follow them. The average user will enjoy Twitter for its freedom of speech, the ability to express whenever they want and for businesses to promote their products and services. Many large corporations are now using Twitter as a platform to obtain information relating to their company, products and services. By leveraging the underlying data, they further strategize and align their products and services with a new target market, or develop new sales strategies for existing ones.

This research focuses on tweets about natural disasters and insurance. Currently, the insurance industry is taking advantage of Twitter, posting links to jobs, posting about customer experiences, information on claims, to name a few. It has allowed those who “follow” them to be more engaged. As Fan and Gordon (2014) discussed, although Twitter has done a lot of good for companies, it has also provided a new set of opportunities and challenges for them. Text analytics - the analysis of unstructured text to find meaning - is one way to deal with the new opportunities and challenges that social media has created. We used the program R to obtain tweets that included the word ‘insurance’ in combination with other natural disaster words (e.g., snow, ice, flood, etc.). In addition, tweets related to 12 insurance companies were collected to identify any pattern about the industry’s products and services as a whole. A total of 11,495 natural disaster tweets and 19,318 insurance company tweets were analyzed using association rule mining.

Association rule mining is one of the major data mining techniques that is used to discover statistically verified co-occurrences of events or transactions in large data sets (Witten & Frank, 2005). Association discovery has been applied in many contexts with the most common application being in market-basket analysis where analysts attempt to find items that are often bought together (Witten & Frank, 2005). The discovery of strong rules in this context suggests product display or service configuration strategies for businesses. It also suggests product and/or service bundling strategies for businesses based on customer past purchase patterns. In this study, we applied association discovery of key terms related to insurance and natural disaster in users’ tweets related to six top Canadian insurance companies as well as the top five insurance companies from the rest of the world, including the new entrant Google Insurance.

According to the results of our analysis, the strongest rule showed that 100% of tweets that contain the words “Earthquake” and “Policy” or “Policyholder” will also include any of the following words: “Home”, “House”, “Homeowners”, “Family”, “Pipes”, “Roof”, and “Property”. This rule occurred 1150 times in the natural disaster tweet which implies that people who tweet about earthquake insurance policies are very concerned with houses. The second strongest rule showed that 92% of tweets that contain the words “Claim” or “Cover” or “Covered” or “Recovery” and “Fire” or “Burn” will also contain the words “Wind” or “Hurricane” or “Hail” or “Storm” and “Theft” or “Loss”. This implies that people who tweet about a claim or being covered desire coverage that includes fire, wind, hurricane, hail, storms, theft and loss. These findings suggest product bundling preferences by customers with potential implications for premiums and other service related issues. Our research also shows the potential text mining applications offer for insurance companies in designing their products and services.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 5: 4 Issues (2018): 1 Released, 3 Forthcoming
Volume 4: 4 Issues (2017)
Volume 3: 4 Issues (2016)
Volume 2: 4 Issues (2015)
Volume 1: 4 Issues (2014)
View Complete Journal Contents Listing