The data points in the graphs should have error bars. If you take multiple samples then average them, the random error will tend to cancel out. Having error b

The data points in the graphs should have error bars. If you take multiple samples then average them, the random error will tend to cancel out. Having error bars would make it clearer how significant the time of posting really is.