Processed Version of Copenhagen Networks Study (CNS) Dataset

The CNS social network of interactions was used in the following manner. To consider close encounters, the data was mined taking into account that the probability for infection is inversely correlated with distance, and decreases dramatically as the distance. The proximity information was registered as the Received Signal Strength (RSSI) function. However, it is hard to infer exact distance information from the RSSI. Moreover, the distance itself is not a single component in calculating the chance of infection - the directions of standing, ventilation, and the environment are also significant parameters. To simplify the model we do not consider all the above. As stronger signal correlates roughly with high proximity, we consider interactions for the RSSI >= -90. We further mined the dataset for cores, and incorporated core information into our modeling. Then consecutive meetings were aggregated. Note that the "datetime" field is arbitrary.