It seems I got my wires cross when I was working on this script. It looks like I
was mixing two aims up. The first half of the script wanted to compare the two
Light Meter reading from the same day. The second half wanted to compare the
same Light Meter (I.E. just Light Meter 1) on the 'same day' across June and
July. In other words, the second half want to compare the readings taken by
Light Meter 1 on the 17th June and July, for example. I guess this is what
happens when you're programming after not sleeping for close to fourty-eight
hours.
Because this script could go in one of two ways, I decided to stick with the
comparing both Meters on the same day -- compare Meter 1 and 2 on the 16th June,
as an example. The changes in this commit reflect this decision.
I replaced it with assets/side-by-side-comparison-2021-07-19.png. This is
prep. work for the bug fixes I need to apply to the day-to-day-comparisons.py
script and doc's.
I've stubbed out each file for this commit. I need to go in and write the main
content for them all (apart from sql-statements.org). I was originally going to
put a breakdown of what each (Python) script does in the wiki but I changed my
mind. The wiki, used by Gitea, doesn't allow .org files and I quite like the
option to run code in any of the (doc) files if needed (using org-babbel).
I need to update README.org to reflect this change.
Because the /output is an empty directory (on a fresh clone of the repo.), the
Python scripts don't have a place to save the charts. This addition just makes
sure the code runs smoother on freshly cloned repository.
This script takes the 'same' day from each month (E.G. the 16th June and July)
and overlays each others charts on top of each other.
If there is no data for both days (E.G. data for the 15th of June but not for
the 15th July), no chart is produced.
Because each chart has inconsistent timestamps (between the two days), each
reading is given a 'Reading Id.' so the time element of each chart is lost: THEY
ARE NOT ONE-TO-ONE MAPPING OF EACH DAY. The charts produced with this script
produce charts depicting the amount of change as a set of sequential events and
that's it. They show a general pattern to the day. For example, a chart will
show an increase in light from point A to Point B for Light Meter 1. It will not
specify how long it took for the change to occur.
This script places all the daily reading for a given Light Meter and overlays
them on top of each other on the chart. Because each file (for each day)
have different combinations of date/timestamps, the charts has to switch to
using plain integers for the x-axis. This means you can't glean much data from a
time/date spective but the charts for both Light Meters do help relay the
amount of activity which happened in them.
I didn't mean to create this script.
It takes the individual daily readings for a Light Meter and places each 'daily
readings' next to each other, on the same graph. This is essentially stitching
the full log of reading back together but with gaps between the individual
charts. The gaps are formed, I think -- I haven't checked -- because the
'missing data' (when the factory was closed or no welding was happening) is
being treated as such. I'm assuming it's because I've passed each 'daily line'
as separate lines so Bokeh has no reason to join them together.
This script makes a HTML chart for both Light Meters. And, the charts are slow
and not very nice to deal with. I've only kept the script and the charts it
produces because the charts look interesting as images and I want to see what
Nic thinks. She might find them useful from an artistic perspective.
The script now processes all the daily files for both Light Meters. Each day,
for each Light Meter, has their own chart (HTML) made as the output of the
script.
The script can process on file and create plot (HTML file) but the aim is to
have the script cycle through all the daily files (for Light Meter 1) and
produce charts of each one.
This script takes all the individual files in /data/light-meter-1-hourly-totals
and creates a line chart from the data. Each line tracks the number of readings
taken in a given hour and each line represents one day during the exhibitions
run time (13th June 2021 to 1st August 2021). The lines are layered on top of
each other in a single grid.
This script produces a line chart, using Bokeh, displaying the daily totals for
both Light Meters. The data is process from /data and the chart is written to
the /output directory -- as a HTML file.
I added checks to stop this script from processing data which was not recorded
during the exhibition's opening times.
Note: The exhibition officially closes on 1st August 2021 but that was a Sunday
and no one was working in Ritherdon that day. This is why the looping range has
been reduced from 6-8 to 6-7.
The exhibition officially ended 1st August 2021 but no one was working in
Ritherdon on that day (Sunday). So, there was not light readings taken. This
script originally was extracting data all the way up to the 31st August
2021. This meant the old version of this script was extracting excess data and
producing files with irrelevant data. The changes here remove the excess data
extraction and processing.