Zooming in on the daily keograms, the motion of celestial objects becomes clear. These daily keograms from January 16th to February 2nd 2024 show the Moon moving across the sky. It starts at sunset near first quarter, culminates around full moon, and ends at sunrise near last quarter.

@cgbassa
I was fascinated of that bright ball during the night that suddenly disappeared in the middle of the sky when ot became day. It's the moon! And it faded away because everything else became brighter at sunrise. That is so cool to watch!
@limebar

@cgbassa very nice. Thanks for sharing!

@cgbassa why do the red areas/curves have such a kink in them?

@cgbassa are the times "true" local times (1200 = sun is the highest) and also not considering DST as far as I can tell?

@flo You are not the first to see the resemblance 😆

@flo @cgbassa I thought that, too!

@CelloMomOnCars @cgbassa my understanding is that it's mostly about the change in solar noon (from the varying speed of the earth along its orbit around the sun) partly compensating the change in day length. https://www.timeanddate.com/astronomy/equation-of-time.html has a more technical explanation

@coldclimate @cgbassa was wondering the same, it's such a cool data visualisation exercise!

@coldclimate Not in detail as many others before me have written about designing/building all-sky cameras and the software needed to capture individual images and create keograms, see https://github.com/AllskyTeam/allsky/. I've only taken their work a small step further. My main changes have been to the capture software to change the exposure logic to also be able to capture during daytime. All individual exposures are offloaded to storage on another computer. On that I create the daily keograms, which are then combined into these year-long keograms, all with custom python code.

@tero That's correct. The local time is essentially kept at CET throughout the year for this plot.

@chrysn The yellow is due to low-altitude cloud cover and fog, which, with this camera, comes out as yellowish, probably due to illumination from streetlights. There is no dimming of street lights, more that clouds change color to gray/white in daytime. Here's a plot of all of December 2024.

@wolf480pl You are not the first to make that connection! In this case the sky is blue during the day and black at night (i.e. not gold and white).

@Dubikan @cgbassa it's the Netherlands, so sunset the 21st of July is around 21:45 and sunrise around 5:45 (we're very far West in our timezone, worsened by daylight saving time)

@xxx This is due to the approach I've chosen for the exposure settings. The camera has controls for exposure and gain, and I've implemented the exposure logic such that from daytime to night time, the gain stays at the daytime value until the exposure time increases to 15 seconds. After that, the gain gets to increase to a maximum value. At Sun rise, the opposite happens (gain deceases first, then exposure time). This plot shows the evolution of exposure time and gain during 24h and can be compared to the keogram and the times of Sun set and Sun rise (red vertical lines). It's clear that the transition from in color coincides with the gain change, which occurs in deep twilight.

@cosmicspittle I based the design of my camera on plans from Thomas Jacquin (see https://www.thomasjacquin.com/make-your-own-allsky-camera/ and https://www.instructables.com/Wireless-All-Sky-Camera/). He also started the github repo for software to run on the RPi. That repo has grown quite a bit and now supports other cameras as well. See https://github.com/AllskyTeam/allsky/. It took me a few months to get the system properly weather proofed.