• ChaoticNeutralCzech@feddit.org
    link
    fedilink
    English
    arrow-up
    8
    ·
    edit-2
    4 months ago

    The color scale hugely amplifies minor differences, see my other comment. In the dataset, 15 years (2000-2014) are represented and the weekly cycle is therefore present. This could have been mitigated by using a 28-year dataset. Here is how often each month started on a given day in the dataset. We don’t have colored text so I used emoji.

    01/ JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
    SUN 2➖ 2➖ 1🔻 3⚠️ 2➖ 3⚠️ 3⚠️ 2➖ 2➖ 2➖ 1🔻 2➖
    MON 2➖ 1🔻 2➖ 2➖ 2➖ 1🔻 2➖ 2➖ 3⚠️ 3⚠️ 2➖ 3⚠️
    TUE 3⚠️ 3⚠️ 2➖ 3⚠️ 3⚠️ 2➖ 3⚠️ 2➖ 1🔻 2➖ 2➖ 1🔻
    WED 2➖ 2➖ 2➖ 1🔻 2➖ 2➖ 1🔻 3⚠️ 2➖ 3⚠️ 2➖ 2➖
    THU 2➖ 2➖ 3⚠️ 2➖ 3⚠️ 2➖ 2➖ 2➖ 2➖ 1🔻 3⚠️ 2➖
    FRI 1🔻 3⚠️ 2➖ 2➖ 1🔻 3⚠️ 2➖ 3⚠️ 2➖ 2➖ 2➖ 2➖
    SAT 3⚠️ 2➖ 3⚠️ 2➖ 2➖ 2➖ 2➖ 1🔻 3⚠️ 2➖ 3⚠️ 3⚠️

    Here is how likely any date is to be Monday-Friday in the data (out of 15):

    How often Mon-Fri JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
    01/08/15/22/29 10🔻 11➖ 11➖ 10🔻 11➖ 10🔻 10🔻 12⚠️ 10🔻 11➖ 11➖ 10🔻
    02/09/16/23/30 11➖ 12⚠️ 12⚠️ 10🔻 11➖ 11➖ 10🔻 11➖ 10🔻 10🔻 12⚠️ 10🔻
    03/10/17/24/31 10🔻 11➖ 11➖ 10🔻 10🔻 12⚠️ 10🔻 11➖ 11➖ 10🔻 11➖ 11➖
    04/11/18/25 10🔻 10🔻 11➖ 11➖ 10🔻 11➖ 11➖ 10🔻 12⚠️ 10🔻 11➖ 12⚠️
    05/12/19/26 11➖ 11➖ 10🔻 12⚠️ 10🔻 11➖ 12⚠️ 10🔻 11➖ 11➖ 10🔻 11➖
    06/13/20/27 12⚠️ 10🔻 10🔻 11➖ 11➖ 10🔻 11➖ 10🔻 11➖ 12⚠️ 10🔻 11➖
    07/14/21/28 11➖ 10🔻 10🔻 11➖ 12⚠️ 10🔻 11➖ 11➖ 10🔻 11➖ 10🔻 10🔻

    You can see for example that August 01/08/15/22/29 is 10-20% more likely to be a workday than other days, which corresponds to the extra births. Feb 29 cannot be read directly from this table; its value is 3 out of 4 (2000: Tue, 2004: Sun, 2008: Fri, 2012: Wed).

    As for your question: every month has approximately the same strength of the weekly cycle, it’s just that the prevailing colors in July-September show them the most. August’s is more visible because it does not have the disrupting July 4 and September 11.