This has frustrated me as a python user for the last 7 years, working as the only python user in business environments dominated by Excel. People will say things like, "if you leave, who can support this report you made in python?" Well I say, who can support the bloated 40mb spreadsheet that would take forever to unpick and figure out how to update with new data? No one can, because I've seen people would rather rebuild their own spreadsheet from scratch, than use the files they inherited from the last person.

If these tools are necessary to conduct business and they are so worried about being able to support it, why don't they use proper software for that process?

A lot of people who make these bloated spreadsheets are people with no education in computing, and don't think about the basics of how to store data that is easy to analyse later. If they are building a weekly report, they build the report and enter the data directly into the report structure, which then makes it almost impossible to analyse later. Next week they just copy the file, rename it and update the data. If you want to analyse that same data over a year, good luck! You can't even count on the data being in the same place over the 52 weeks, since they would have added and removed data points over time.

Once I got the process down in a jupyter notebook, handling all the oddities with the data coming from whichever website, CSV file, data warehouse report I need, I can just save it as a .py file and run it as a scheduled task on a virtual computer forever. The data is kept in a format that can be appended to with each update, and can be easily analysed later.

The most amazing thing with replacing excel with python is you don't need to manually perform the update process yourself. Which means it doesn't cost anything to run the process more often. Weekly reports can become daily, or even hourly email updates that are only sent when something interesting happens. People can start reacting to things shortly after they happen, rather than having to remember what happened a week or a month ago. The iteration on improving becomes so much faster. People spend more of their time discussing how to fix problems, rather than spending time building problem finders. You can even start to automate the fixing of the problem in python and people don't even have to spend time on that thing at all, ever again.