Pufferfish privacy when publishing on thematic maps

Statistical disclosure control is essential to ensure the privacy of individuals in a data set. Statistics of a data set can be visualized on thematic maps by colouring its geographic location. We will apply Pufferfish privacy to protect thematic grid maps by adding appropriate noise. We come up with absolute and relative error protection methods by applying the least necessary noise according to the Laplace distribution. This way, the highest utility remains, while privacy is guaranteed. After observing some unrealistic results, we have also developed a bounded privacy mechanism. Numerical experiments show how the mechanisms act on different input parameters.