I went through a phase of trying to automate _all_ aspects of my life. If I did a task more than 3 times, and it could be automated, I went ahead and wrote some scripts for that.
Anyone will quickly learn there is a trade-off between managing the time it takes to maintain automation and the time you would spend just "doing" said task.
After this realization, I started manually performing tasks, timing them, and storing them in a glorified spreadsheet. Any time automation broke, I timed how long it took to fix, and also put it in the spreadsheet.
I deprecated all automation tasks that did not save me either 10% or more time, or tasks that I found at least some enjoyment in (such as messaging my wife when I would be home, reading my child's school reports regardless of their result).
I have over 140 automation scripts in tact. Some are software-only (interacting with APIs, emails, SMS, scraping) and some have associated hardware (sensors attached, etc).
I have become obsessed with data being generated by this automation, so I now log all events and meta information where possible. This gives me pretty good insight on where pointless items exist in my life (which allows me to just stop doing them, automated or manually) and where important parts also exist.
Would you mind sharing which automated tasks offered the highest ROI, and some information about how you run them all (i.e., do you use cloud services, or an always on RPi that you run the scripts on)? Also, which libraries have you used for scraping?
Help me understand the rationale behind the original automation effort on reading your child's school reports? I would have thought that is an event that is pretty infrequent, quite important, with not enough historical data to train a useful model.
Anyone will quickly learn there is a trade-off between managing the time it takes to maintain automation and the time you would spend just "doing" said task.
After this realization, I started manually performing tasks, timing them, and storing them in a glorified spreadsheet. Any time automation broke, I timed how long it took to fix, and also put it in the spreadsheet.
I deprecated all automation tasks that did not save me either 10% or more time, or tasks that I found at least some enjoyment in (such as messaging my wife when I would be home, reading my child's school reports regardless of their result).
I have over 140 automation scripts in tact. Some are software-only (interacting with APIs, emails, SMS, scraping) and some have associated hardware (sensors attached, etc).
I have become obsessed with data being generated by this automation, so I now log all events and meta information where possible. This gives me pretty good insight on where pointless items exist in my life (which allows me to just stop doing them, automated or manually) and where important parts also exist.