As a financially avid reader, I am sure that you can roughly say how much money you need for a month. You can assist your gut feeling with tools like Kmymoney, Homebank, GnuCash et al. to know exactly how much it is that you spend and earn per month. However, the aforementioned tools have a common shortcoming, at least in my opinion: They use different data formats to store data. If the software reaches end of life, you could lose months or even years of transactional data because you might not be able to extract it from the old storage format.
Enter Plain Text Accounting (PTA). A text file will store your transactions using a well-defined format. No matter what happens to the analytics software, you will always be able to manipulate and convert from text - solving the problem above.
PTA tools tend to be designed for the command line, which can be somewhat counter-intuitive if you expect graphs and visualizations of the GUI tools. In this article I will show you how to easily work with beancount and visualize your numbers nicely.
First, you install beancount from any source that works for your system.
Second, create a file and folder structure like so:
main.beancount and open some accounts. You can use this article as a guide. It doesn’t have to be perfect from the start, you can always retry.
Fourth, create folders for your bank accounts inside
raw and create empty beanfiles for them.
Fifth, use the
include command inside
main.beancount to pull in the account-specific beanfiles your just created.
If you’re like me, you have two accounts where money shuttle back and forth. Then you would have a folder structure similar to the following:
As mentioned in the introduction, I perceive the big strength of PTA that you can process anything. For this job, beancount offers a so-called Importer API. You can either write your own Python importer or copy and expand on one from Github. I used the following two in my projects:
Either install the importer via
pip or make it importable for Python from the
importers/ directory. It will process your input files line-by-line and create the appropriate accountin entries.
After having found and configured an importer, you can wire it up quite easily in
import.config. Here is how it looks for me:
This tells beancount the following: “Please use these two importers when importing transactions. The importers will tell you whether they are capable of importing a given file, so the right one will get selected automatically.”
A month has passed while you tried to set this up, so now you have a month’s worth of data ;-)
Download it as CSV, and save it under
raw/n26-01-21.csv. Then, you can append the plain text ledger to the
I recommend checking the output in the beginning, as categories could be wrong. Maybe you notice some patterns that can help you automatically categorize transactions inside the importer. I did that and it saves quite a bit of work.
Here is why I picked beancount: fava. It’s a great visualization frontend for beancount ledgers. After adding a month’s worth of data, I just fire it up and can analyze what happened.
If you’d like to do so, too, simply install it and run: