TY - JOUR

T1 - FreeTxt

T2 - A corpus-based bilingual free-text survey and questionnaire data analysis toolkit

AU - Knight, D.

AU - Khallaf, N.

AU - Rayson, P.

AU - El-Haj, M.

AU - Ezeani, I.

AU - Morris, S.

PY - 2024/12/31

Y1 - 2024/12/31

N2 - Qualitative free-text responses (e.g. from questionnaires and surveys) pose a challenge to many companies and institutions which lack the expertise to analyse such data with ease. While a range of sophisticated tools for the analysis of text do exist, these are often expensive, difficult to use and/or inaccessible to non-expert users. These tools also lack support for the analysis of English and Welsh text, which can be a particular challenge in the bilingual context of Wales. This paper details the key functionalities of the first corpus-based ‘FreeTxt’ toolkit which has been designed to support the systematic analysis and visualisation of free-text data, as a direct response to these two key needs. This paper demonstrates how, by working in partnership, software engineers, natural language processing (NLP) experts and corpus linguists can collaborate with end-users and beneficiaries to provide effective solutions to real world problems. Through the development of FreeTxt (www.freetxt.app), we aimed to empower end-users to direct and lead their own analyses of both small-scale and more extensive datasets to maximise the reach and potential impact generated. The approaches reported here, and the bilingual toolkit developed, can be replicated and extended for use in other language contexts and across a range of public and professional sectors. FreeTxt is now available for the analysis of Welsh and/or English, for use by anyone in any sector in Wales and beyond.

AB - Qualitative free-text responses (e.g. from questionnaires and surveys) pose a challenge to many companies and institutions which lack the expertise to analyse such data with ease. While a range of sophisticated tools for the analysis of text do exist, these are often expensive, difficult to use and/or inaccessible to non-expert users. These tools also lack support for the analysis of English and Welsh text, which can be a particular challenge in the bilingual context of Wales. This paper details the key functionalities of the first corpus-based ‘FreeTxt’ toolkit which has been designed to support the systematic analysis and visualisation of free-text data, as a direct response to these two key needs. This paper demonstrates how, by working in partnership, software engineers, natural language processing (NLP) experts and corpus linguists can collaborate with end-users and beneficiaries to provide effective solutions to real world problems. Through the development of FreeTxt (www.freetxt.app), we aimed to empower end-users to direct and lead their own analyses of both small-scale and more extensive datasets to maximise the reach and potential impact generated. The approaches reported here, and the bilingual toolkit developed, can be replicated and extended for use in other language contexts and across a range of public and professional sectors. FreeTxt is now available for the analysis of Welsh and/or English, for use by anyone in any sector in Wales and beyond.

U2 - 10.1016/j.acorp.2024.100103

DO - 10.1016/j.acorp.2024.100103

M3 - Journal article

VL - 4

JO - Applied Corpus Linguistics

JF - Applied Corpus Linguistics

SN - 2666-7991

IS - 3

M1 - 100103

ER -