Multivariate statistical analysis of thematic changes in customer feedback

Tässä opinnäytetyössä tehdään asiakaspalautteen teemamuutosten tilastollinen monimuuttuja-analyysi keskittyen ensisijaisesti monimuuttujamenetelmiin. Tutkimusaineisto on hankittu Aiwo Digital Oy:ltä, joka on saanut aineiston asiakasyrityksiltään. Analyysi keskittyi pseudonymisoituihin teemamuuttujii...

Täydet tiedot

Bibliografiset tiedot
Päätekijä: Lopperi, Mikko
Muut tekijät: Matemaattis-luonnontieteellinen tiedekunta, Faculty of Sciences, Matematiikan ja tilastotieteen laitos, Department of Mathematics and Statistics, Jyväskylän yliopisto, University of Jyväskylä
Aineistotyyppi: Pro gradu
Kieli:eng
Julkaistu: 2023
Aiheet:
Linkit: https://jyx.jyu.fi/handle/123456789/88486
_version_ 1828193042778030080
author Lopperi, Mikko
author2 Matemaattis-luonnontieteellinen tiedekunta Faculty of Sciences Matematiikan ja tilastotieteen laitos Department of Mathematics and Statistics Jyväskylän yliopisto University of Jyväskylä
author_facet Lopperi, Mikko Matemaattis-luonnontieteellinen tiedekunta Faculty of Sciences Matematiikan ja tilastotieteen laitos Department of Mathematics and Statistics Jyväskylän yliopisto University of Jyväskylä Lopperi, Mikko Matemaattis-luonnontieteellinen tiedekunta Faculty of Sciences Matematiikan ja tilastotieteen laitos Department of Mathematics and Statistics Jyväskylän yliopisto University of Jyväskylä
author_sort Lopperi, Mikko
datasource_str_mv jyx
description Tässä opinnäytetyössä tehdään asiakaspalautteen teemamuutosten tilastollinen monimuuttuja-analyysi keskittyen ensisijaisesti monimuuttujamenetelmiin. Tutkimusaineisto on hankittu Aiwo Digital Oy:ltä, joka on saanut aineiston asiakasyrityksiltään. Analyysi keskittyi pseudonymisoituihin teemamuuttujiin, jotka ovat binäärikoodattuja, ja osoittavat, esiintyikö teema yksittäisessä palautteessa. Teemojen lisäksi datassa oli taustamuuttujia ja tunne, joka ilmaisi palautteen sävyä. Ensisijaisena tavoitteena oli ryhmitellä teemat, jotka käyttäytyivät samalla tavalla tutkimusjakson aikana. Käytimme hierarkkista ryhmittelyä binäärisen monimuuttujadatan ryhmittelemiseen. Opinnäytetyössä tarkastellaan erilaisia samanlaisuusmittoja binääristen teemavektoreiden välillä ja erilaisuusmittoja ryhmien välillä. Aukkosuuretta ja siluettisuuretta tarkasteltiin kriteereinä optimaalisen ryhmämäärän valintaan. Ryhmittelimme 79 teemamuuttujaa kahteen ryhmään. Aggregoimme päivittäisen datan viikkotasolle ja tutkimme eri teemaryhmien teemaesiintymiä. Löysimme seitsemän teemaa (ryhmä 1), jotka osoittivat samanlaista käyttäytymistä koko tutkimusjakson ajan. Käsittelimme metrisen moniulotteisen skaalauksen (MDS) teoriaa ja käytimme MDS:ää moniulotteisen teemadatan visualisointiin matalaulotteisessa avaruudessa. Laskimme uusioluottamusvälit teemaesiintymille. Tutkimalla luottamusvälejä havaitsimme, että kaikki ryhmän 1 muutokset eivät näyttäneet johtuvan ainoastaan satunnaisesta vaihtelusta. Käytimme negatiivista binomiregressiota temaesiintymien mallintamiseen viikosta ja tunteesta riippuen. Palaute, jossa ryhmän 1 teemoja esiintyi, oli enimmäkseen negatiivista. Tulosten tulkintaa varten, saimme Aiwolta jälkikäteen tiedot todellisista teemoista, jotka olivat ryhmän 1 pseudonymisoitujen teemojen takana. Viisi teemaa ryhmässä 1 liittyi käytettävyyteen ja kaksi asiakaspalveluun. Päättelimme, että näiden teemojen muutokset saattoivat johtua käyttöliittymän tai joidenkin asiakassovellusten käyttötavan muutoksesta. Negatiivinen palaute voi indikoida, miten käytettävyyden muutokset on otettu vastaan. On syytä myös huomioida, että merkittävä määrä negatiivista palautetta annetaan tyypillisesti silloin, kun jokin ei toimi odotetulla tavalla. Tarkempi tulkinta vaatisi asiakaspalautteiden analysointia tekstitasolla tai asiakasyrityksen omaa arviota. This thesis conducts a multivariate statistical analysis of thematic changes in customer feedback, primarily focusing on multivariate methods. The study data were obtained from Aiwo Digital Oy, which received it from their client companies. The analysis focused on pseudonymized binary-coded theme variables, which indicate whether the theme occurred in an individual feedback. In addition to themes, there were also background variables, and sentiment, which indicated the tone of the feedback. The primary goal was to group themes that behaved similarly over the study period. We applied hierarchical clustering to group the binary multivariate data. The thesis discusses various similarity measures between binary theme vectors and dissimilarity measures between clusters. The gap statistic and the silhouette coefficient were considered criteria for choosing an appropriate number of clusters. We clustered 79 theme variables into two groups. We aggregated data on a weekly basis and investigated the theme occurrences of different theme groups. Finally, we discovered seven themes (Group 1) that exhibited similar behavior throughout the study period. We discussed the theory of metric multidimensional scaling (MDS) and applied metric MDS to visualize the multidimensional theme data in a low-dimensional space. We calculated bootstrap confidence intervals for theme occurrences. Through an investigation of the confidence intervals, we discovered that not all changes in Group 1 appeared to be solely due to natural variation in the data. We applied negative binomial regression to model theme counts depending on the week and the sentiment. Feedback in which themes of Group 1 occurred appeared to be primarily negative. For an interpretation of the results, after the study, we were given the real themes behind the pseudonymized themes of Group 1 by Aiwo. Five themes in Group 1 related to usability and two to customer service. We concluded that the changes in these themes were likely due to the change in the user interface or in the method of use of some client applications. The negative feedback may indicate how the changes in usability have been received. Still, it is also worth noting that negative feedback is typically received when something does not function as expected. A thorough analysis of the customer feedback at the text level or the client's assessment would be necessary for a more accurate interpretation.
first_indexed 2023-08-02T20:20:42Z
format Pro gradu
free_online_boolean 1
fullrecord [{"key": "dc.contributor.advisor", "value": "Vihola, Matti", "language": "", "element": "contributor", "qualifier": "advisor", "schema": "dc"}, {"key": "dc.contributor.author", "value": "Lopperi, Mikko", "language": "", "element": "contributor", "qualifier": "author", "schema": "dc"}, {"key": "dc.date.accessioned", "value": "2023-08-02T09:03:37Z", "language": null, "element": "date", "qualifier": "accessioned", "schema": "dc"}, {"key": "dc.date.available", "value": "2023-08-02T09:03:37Z", "language": null, "element": "date", "qualifier": "available", "schema": "dc"}, {"key": "dc.date.issued", "value": "2023", "language": "", "element": "date", "qualifier": "issued", "schema": "dc"}, {"key": "dc.identifier.uri", "value": "https://jyx.jyu.fi/handle/123456789/88486", "language": null, "element": "identifier", "qualifier": "uri", "schema": "dc"}, {"key": "dc.description.abstract", "value": "T\u00e4ss\u00e4 opinn\u00e4ytety\u00f6ss\u00e4 tehd\u00e4\u00e4n asiakaspalautteen teemamuutosten tilastollinen monimuuttuja-analyysi keskittyen ensisijaisesti monimuuttujamenetelmiin. Tutkimusaineisto on hankittu Aiwo Digital Oy:lt\u00e4, joka on saanut aineiston asiakasyrityksilt\u00e4\u00e4n. Analyysi keskittyi pseudonymisoituihin teemamuuttujiin, jotka ovat bin\u00e4\u00e4rikoodattuja, ja osoittavat, esiintyik\u00f6 teema yksitt\u00e4isess\u00e4 palautteessa. Teemojen lis\u00e4ksi datassa oli taustamuuttujia ja tunne, joka ilmaisi palautteen s\u00e4vy\u00e4.\n\nEnsisijaisena tavoitteena oli ryhmitell\u00e4 teemat, jotka k\u00e4ytt\u00e4ytyiv\u00e4t samalla tavalla tutkimusjakson aikana. K\u00e4ytimme hierarkkista ryhmittely\u00e4 bin\u00e4\u00e4risen monimuuttujadatan ryhmittelemiseen. Opinn\u00e4ytety\u00f6ss\u00e4 tarkastellaan erilaisia samanlaisuusmittoja bin\u00e4\u00e4risten teemavektoreiden v\u00e4lill\u00e4 ja erilaisuusmittoja ryhmien v\u00e4lill\u00e4. Aukkosuuretta ja siluettisuuretta tarkasteltiin kriteerein\u00e4 optimaalisen ryhm\u00e4m\u00e4\u00e4r\u00e4n valintaan. Ryhmittelimme 79 teemamuuttujaa kahteen ryhm\u00e4\u00e4n. Aggregoimme p\u00e4ivitt\u00e4isen datan viikkotasolle ja tutkimme eri teemaryhmien teemaesiintymi\u00e4. L\u00f6ysimme seitsem\u00e4n teemaa (ryhm\u00e4 1), jotka osoittivat samanlaista k\u00e4ytt\u00e4ytymist\u00e4 koko tutkimusjakson ajan.\n\nK\u00e4sittelimme metrisen moniulotteisen skaalauksen (MDS) teoriaa ja k\u00e4ytimme MDS:\u00e4\u00e4 moniulotteisen teemadatan visualisointiin matalaulotteisessa avaruudessa. Laskimme uusioluottamusv\u00e4lit teemaesiintymille. Tutkimalla luottamusv\u00e4lej\u00e4 havaitsimme, ett\u00e4 kaikki ryhm\u00e4n 1 muutokset eiv\u00e4t n\u00e4ytt\u00e4neet johtuvan ainoastaan satunnaisesta vaihtelusta. K\u00e4ytimme negatiivista binomiregressiota temaesiintymien mallintamiseen viikosta ja tunteesta riippuen. Palaute, jossa ryhm\u00e4n 1 teemoja esiintyi, oli enimm\u00e4kseen negatiivista.\n\nTulosten tulkintaa varten, saimme Aiwolta j\u00e4lkik\u00e4teen tiedot todellisista teemoista, jotka olivat ryhm\u00e4n 1 pseudonymisoitujen teemojen takana. Viisi teemaa ryhm\u00e4ss\u00e4 1 liittyi k\u00e4ytett\u00e4vyyteen ja kaksi asiakaspalveluun. P\u00e4\u00e4ttelimme, ett\u00e4 n\u00e4iden teemojen muutokset saattoivat johtua k\u00e4ytt\u00f6liittym\u00e4n tai joidenkin asiakassovellusten k\u00e4ytt\u00f6tavan muutoksesta. Negatiivinen palaute voi indikoida, miten k\u00e4ytett\u00e4vyyden muutokset on otettu vastaan. On syyt\u00e4 my\u00f6s huomioida, ett\u00e4 merkitt\u00e4v\u00e4 m\u00e4\u00e4r\u00e4 negatiivista palautetta annetaan tyypillisesti silloin, kun jokin ei toimi odotetulla tavalla. Tarkempi tulkinta vaatisi asiakaspalautteiden analysointia tekstitasolla tai asiakasyrityksen omaa arviota.", "language": "fi", "element": "description", "qualifier": "abstract", "schema": "dc"}, {"key": "dc.description.abstract", "value": "This thesis conducts a multivariate statistical analysis of thematic changes in customer feedback, primarily focusing on multivariate methods. The study data were obtained from Aiwo Digital Oy, which received it from their client companies. The analysis focused on pseudonymized binary-coded theme variables, which indicate whether the theme occurred in an individual feedback. In addition to themes, there were also background variables, and sentiment, which indicated the tone of the feedback.\n\nThe primary goal was to group themes that behaved similarly over the study period. We applied hierarchical clustering to group the binary multivariate data. The thesis discusses various similarity measures between binary theme vectors and dissimilarity measures between clusters. The gap statistic and the silhouette coefficient were considered criteria for choosing an appropriate number of clusters. We clustered 79 theme variables into two groups. We aggregated data on a weekly basis and investigated the theme occurrences of different theme groups. Finally, we discovered seven themes (Group 1) that exhibited similar behavior throughout the study period.\n\nWe discussed the theory of metric multidimensional scaling (MDS) and applied metric MDS to visualize the multidimensional theme data in a low-dimensional space. We calculated bootstrap confidence intervals for theme occurrences. Through an investigation of the confidence intervals, we discovered that not all changes in Group 1 appeared to be solely due to natural variation in the data. We applied negative binomial regression to model theme counts depending on the week and the sentiment. Feedback in which themes of Group 1 occurred appeared to be primarily negative.\n\nFor an interpretation of the results, after the study, we were given the real themes behind the pseudonymized themes of Group 1 by Aiwo. Five themes in Group 1 related to usability and two to customer service. We concluded that the changes in these themes were likely due to the change in the user interface or in the method of use of some client applications. The negative feedback may indicate how the changes in usability have been received. Still, it is also worth noting that negative feedback is typically received when something does not function as expected. A thorough analysis of the customer feedback at the text level or the client's assessment would be necessary for a more accurate interpretation.", "language": "en", "element": "description", "qualifier": "abstract", "schema": "dc"}, {"key": "dc.description.provenance", "value": "Submitted by Paivi Vuorio (paelvuor@jyu.fi) on 2023-08-02T09:03:37Z\nNo. of bitstreams: 0", "language": "en", "element": "description", "qualifier": "provenance", "schema": "dc"}, {"key": "dc.description.provenance", "value": "Made available in DSpace on 2023-08-02T09:03:37Z (GMT). No. of bitstreams: 0\n Previous issue date: 2023", "language": "en", "element": "description", "qualifier": "provenance", "schema": "dc"}, {"key": "dc.format.extent", "value": "65", "language": "", "element": "format", "qualifier": "extent", "schema": "dc"}, {"key": "dc.language.iso", "value": "eng", "language": null, "element": "language", "qualifier": "iso", "schema": "dc"}, {"key": "dc.rights", "value": "In Copyright", "language": null, "element": "rights", "qualifier": null, "schema": "dc"}, {"key": "dc.subject.other", "value": "themes of customer feedback", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "binary data", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "hierarchical clustering", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "the gap statistic", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "the silhouette coefficient", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "multidimensional scaling", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "bootstrap confidence intervals", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.subject.other", "value": "negative binomial regression", "language": "", "element": "subject", "qualifier": "other", "schema": "dc"}, {"key": "dc.title", "value": "Multivariate statistical analysis of thematic changes in customer feedback", "language": "", "element": "title", "qualifier": null, "schema": "dc"}, {"key": "dc.type", "value": "master thesis", "language": null, "element": "type", "qualifier": null, "schema": "dc"}, {"key": "dc.identifier.urn", "value": "URN:NBN:fi:jyu-202308024601", "language": "", "element": "identifier", "qualifier": "urn", "schema": "dc"}, {"key": "dc.type.ontasot", "value": "Master\u2019s thesis", "language": "en", "element": "type", "qualifier": "ontasot", "schema": "dc"}, {"key": "dc.type.ontasot", "value": "Pro gradu -tutkielma", "language": "fi", "element": "type", "qualifier": "ontasot", "schema": "dc"}, {"key": "dc.contributor.faculty", "value": "Matemaattis-luonnontieteellinen tiedekunta", "language": "fi", "element": "contributor", "qualifier": "faculty", "schema": "dc"}, {"key": "dc.contributor.faculty", "value": "Faculty of Sciences", "language": "en", "element": "contributor", "qualifier": "faculty", "schema": "dc"}, {"key": "dc.contributor.department", "value": "Matematiikan ja tilastotieteen laitos", "language": "fi", "element": "contributor", "qualifier": "department", "schema": "dc"}, {"key": "dc.contributor.department", "value": "Department of Mathematics and Statistics", "language": "en", "element": "contributor", "qualifier": "department", "schema": "dc"}, {"key": "dc.contributor.organization", "value": "Jyv\u00e4skyl\u00e4n yliopisto", "language": "fi", "element": "contributor", "qualifier": "organization", "schema": "dc"}, {"key": "dc.contributor.organization", "value": "University of Jyv\u00e4skyl\u00e4", "language": "en", "element": "contributor", "qualifier": "organization", "schema": "dc"}, {"key": "dc.subject.discipline", "value": "Tilastotiede", "language": "fi", "element": "subject", "qualifier": "discipline", "schema": "dc"}, {"key": "dc.subject.discipline", "value": "Statistics", "language": "en", "element": "subject", "qualifier": "discipline", "schema": "dc"}, {"key": "yvv.contractresearch.collaborator", "value": "business", "language": "", "element": "contractresearch", "qualifier": "collaborator", "schema": "yvv"}, {"key": "yvv.contractresearch.funding", "value": "0", "language": "", "element": "contractresearch", "qualifier": "funding", "schema": "yvv"}, {"key": "yvv.contractresearch.initiative", "value": "student", "language": "", "element": "contractresearch", "qualifier": "initiative", "schema": "yvv"}, {"key": "dc.type.coar", "value": "http://purl.org/coar/resource_type/c_bdcc", "language": null, "element": "type", "qualifier": "coar", "schema": "dc"}, {"key": "dc.rights.copyright", "value": "\u00a9 The Author(s)", "language": null, "element": "rights", "qualifier": "copyright", "schema": "dc"}, {"key": "dc.rights.accesslevel", "value": "openAccess", "language": null, "element": "rights", "qualifier": "accesslevel", "schema": "dc"}, {"key": "dc.type.publication", "value": "masterThesis", "language": null, "element": "type", "qualifier": "publication", "schema": "dc"}, {"key": "dc.subject.oppiainekoodi", "value": "4043", "language": "", "element": "subject", "qualifier": "oppiainekoodi", "schema": "dc"}, {"key": "dc.subject.yso", "value": "palaute", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "tilastomenetelm\u00e4t", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "monimuuttujamenetelm\u00e4t", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "asiakkaat", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "feedback", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "statistical methods", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "multivariable methods", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.subject.yso", "value": "customers", "language": null, "element": "subject", "qualifier": "yso", "schema": "dc"}, {"key": "dc.rights.url", "value": "https://rightsstatements.org/page/InC/1.0/", "language": null, "element": "rights", "qualifier": "url", "schema": "dc"}]
id jyx.123456789_88486
language eng
last_indexed 2025-03-31T20:01:13Z
main_date 2023-01-01T00:00:00Z
main_date_str 2023
online_boolean 1
online_urls_str_mv {"url":"https:\/\/jyx.jyu.fi\/bitstreams\/f1d9e055-d8bc-42f8-afdf-5020792d4636\/download","text":"URN:NBN:fi:jyu-202308024601.pdf","source":"jyx","mediaType":"application\/pdf"}
publishDate 2023
record_format qdc
source_str_mv jyx
spellingShingle Lopperi, Mikko Multivariate statistical analysis of thematic changes in customer feedback themes of customer feedback binary data hierarchical clustering the gap statistic the silhouette coefficient multidimensional scaling bootstrap confidence intervals negative binomial regression Tilastotiede Statistics 4043 palaute tilastomenetelmät monimuuttujamenetelmät asiakkaat feedback statistical methods multivariable methods customers
title Multivariate statistical analysis of thematic changes in customer feedback
title_full Multivariate statistical analysis of thematic changes in customer feedback
title_fullStr Multivariate statistical analysis of thematic changes in customer feedback Multivariate statistical analysis of thematic changes in customer feedback
title_full_unstemmed Multivariate statistical analysis of thematic changes in customer feedback Multivariate statistical analysis of thematic changes in customer feedback
title_short Multivariate statistical analysis of thematic changes in customer feedback
title_sort multivariate statistical analysis of thematic changes in customer feedback
title_txtP Multivariate statistical analysis of thematic changes in customer feedback
topic themes of customer feedback binary data hierarchical clustering the gap statistic the silhouette coefficient multidimensional scaling bootstrap confidence intervals negative binomial regression Tilastotiede Statistics 4043 palaute tilastomenetelmät monimuuttujamenetelmät asiakkaat feedback statistical methods multivariable methods customers
topic_facet 4043 Statistics Tilastotiede asiakkaat binary data bootstrap confidence intervals customers feedback hierarchical clustering monimuuttujamenetelmät multidimensional scaling multivariable methods negative binomial regression palaute statistical methods the gap statistic the silhouette coefficient themes of customer feedback tilastomenetelmät
url https://jyx.jyu.fi/handle/123456789/88486 http://www.urn.fi/URN:NBN:fi:jyu-202308024601
work_keys_str_mv AT lopperimikko multivariatestatisticalanalysisofthematicchangesincustomerfeedback