University of Twente Student Theses
Can data from Twitter produce useful feature suggestions in the audio streaming service field? - An example of Spotify
Sandberg, Aaron (2022) Can data from Twitter produce useful feature suggestions in the audio streaming service field? - An example of Spotify.
|
PDF
678kB |
| Abstract: | This research paper examined whether relevant feature suggestions could be found in short messages (tweets) retrieved from Twitter and whether a natural language processing algorithm (BERT) could identify these at the example of Spotify. The data was collected via an API using only messages mentioning Spotify’s help account. Then it was filtered further. All 10,000 tweets were labelled whether they contained a need and clustered according to the size of the proposed feature and which idea they contained. Results showed that 5.22% of all tweets contained a need and the algorithm had a f1-score of 53%. The findings suggest that Twitter can be a source of ideas for Spotify. Furthermore, the analysis of the most suggested ideas showed that many of them can be considered valuable. There were ideas of different sizes present, with smaller ones forming the majority. The data was found to be of considerable velocity as users reacted quickly to events and problems. Variability was displayed in the form of trending suggestions. Results suggest that this need mining approach may serve as a tool to more efficiently collect user suggestions compared to standardly used methods. |
| Item Type: | Essay (Bachelor) |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 54 computer science, 85 business administration, organizational science |
| Programme: | International Business Administration BSc (50952) |
| Link to this item: | https://purl.utwente.nl/essays/90911 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page