The day of week probably matters as well. For example, the conferences will always span a weekend, giving a biased sample of days. I think it is a mistake to tr

The day of week probably matters as well. For example, the conferences will always span a weekend, giving a biased sample of days. I think it is a mistake to try interpreting this data without considering it as a timeseries.