Ever since the General Data Protection Regulation (GDPR) came into force, data anonymization has become even more relevant and important than before, especially for companies wanting to share and analyze their data. This thesis will anonymize smart meter data received from Tekniska Verken and evaluate how the anonymization method performs regarding the trade-off between information loss and privacy for the given dataset. The smart meter data consists of both personal data and the energy consumption recorded at a frequency of every hour. In the thesis, the given dataset is clustered with k-means, dividing the data into different groups depending on the daily energy patterns of each household. K-anonymity is then applied to each of the clusters to anonymize the data. The results show that the method used in this study had better privacy than utility. Additionally, when compared to anonymizing the dataset without clustering it, the method provided a better utility with less privacy protection. Further, two datasets were created with a low re-identification risk to comply with the GDPR. However, it is difficult to say whether the two datasets complied with the GDPR, as it is essentially a decision to be made in a court case.
Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:liu-181156 |
Date | January 2021 |
Creators | Thorgren, Elin |
Publisher | Linköpings universitet, Institutionen för datavetenskap |
Source Sets | DiVA Archive at Upsalla University |
Language | English |
Detected Language | English |
Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
Format | application/pdf |
Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0014 seconds