A Data Analytics Approach to Understanding Samaritans’ Helpline Caller Behaviour
The project used machine learning approaches to seek to understand the behaviour of callers to the Samaritans Ireland telephone helpline
The project undertook a review of the scientific literature on helpline caller behaviour analysis and then used machine learning approaches to seek to understand the patterns of behaviour of callers to the Samaritans telephone helpline. The study adds to previous research on crisis lines and help lines, but expands on this research by focusing solely on caller behaviour. The literature review for this study confirms that no research has been published that is specific to caller behaviour analysed from telephone records only. In addition, no research was found using analysis of large data sets over a prolonged period. This research therefore expands upon previous research by analysing a large telephone data sample over a 3-year period. Confidentiality of service users was maintained as the data was processed to ensure information was not identifiable to any individual service user. The primary focus of the data analysis was on service users’ behaviours; specifically, on how and when they contact the Samaritans helpline. The Samaritans call data presents a specific opportunity to examine caller behaviour to the helpline, through analysis of this data. This research aims to expand the current understanding of the behaviour of Samaritans’ service users to understand how the helpline can maximise accessibility and develop the service to better support both existing and new service users in the future. This work indicates the value of data science in understanding the types of customers a service has, their behaviour and the potential use of data for optimising the service as well as developing more in-house data-informed policies and strategies.
Funder: Samaritans Ireland