Open Source #TBT and a side project

Josiah Mugambi

Feb 17, 2022 • 3 min read

This is a bit of a #ThrowBack now that it’s Thursday + a side project.

Once upon a time, I was quite the vociferous open source evangelist and got into lots of heated debates with some then Microsoft evangelists. I contributed a little to the debate in 2007-2008 on the OOXML standard that took place on the Skunkworks mailing list.

Kenya ended up abstaining - https://lnkd.in/grnSHHax, though the standard was approved despite our efforts - https://edri.org/our-work/edrigramnumber6-7ooxml-adopted-iso/ and I shall link to a few more of my other posts below /o\.)

14+ years on:

Microsoft is a leading open source contributor, and held #Azure Open Source Day earlier this week
Lots of open source projects form the kernel of highly profitable cloud, SaaS platforms.
Open source databases overtook closed source databases in popularity in early 2021 (https://db-engines.com/en/ranking_osvsc)

This from earlier this year (via Christer Gundersen on LinkedIn) is a deeper look into open source challengers in other categories:

https://rajko-rad.medium.com/the-rise-of-open-source-challengers-4a3d93932425

This is useful list of open source alternatives:

The decision to go open source, I believe, is now one that is largely driven from a business point of view, and is less ideological. However, it is doubtful that some projects and numerous products we currently use and probably take for granted would exist if it wasn’t for open source. This brings me to my side project; one that I have had on my mind for several years now, but only got round to starting on late last year.

Keeping tabs on your personal finances is a good thing to do. However, most people I know don’t daily track their personal finances. I believe it is primarily because it is a tedious and boring task. So I’ve been thinking about how take away the mundane and tedious aspects using tech, reducing it to a minimum, while further enhancing up my AI/ML and data skills as I create something useful for myself (and maybe for others). Such a digital personal finance assistant is probably less useful if transactions are not done or data available digitally (e.g. via bank, mobile money (such as Safaricom MPESA), or even crypto)

I’ve began working on bits of it sporadically over the last few weeks. It’s still early days and I first trying to automatically classify mobile money transaction extracts in this case my Safaricom’s MPESA statements (while refreshing my engineering brain).

In the spirit of open source, this is a Jupyter notebook what I have done so far: https://github.com/josiahmugambi/experiments-with-data/tree/main/mpesa_pdf_statements and my notes: https://learn.mugambi.co.ke/notes/data_txts.html

I would love any feedback. I’m also keen to hear from anyone those who would find a digital personal finance assistant or bot useful, or have used something similar.

Once upon in .ke Open Source:

https://josiahmugambi.wordpress.com/2007/04/10/free-and-open-source-software-foss-in-kenya/

https://josiahmugambi.wordpress.com/2007/06/05/why-should-open-source-software-be-used-in-schools/

https://josiahmugambi.wordpress.com/2008/03/31/update-ooxml-iso-vote/

Sign up for more like this.