My first drawing was much simpler
When I first tried to explain Culture Atlas, I used five kinds of things: places, languages, communities, traditions and stories.
I imagined them as dots with lines between them. A tradition could connect to a community. A community could connect to a language. A story could connect to a place.
It made sense on paper. Then I started putting real information into the project.
Real culture does not always fit my boxes
I soon needed more kinds of records. Where should a festival go? Is an old temple just a place? What about a historical event, a poem, a deity or a UNESCO World Heritage site?
The database now has more categories than my first drawing. It can hold things such as festivals, people, historical contexts, archaeological sites, cultural texts, media and time periods as well as the original types.
I learned an important lesson: if something does not fit my model, I should question the model before I force the information into the wrong box.
So what is a data model?
I kept seeing the words data model while building this. The simplest way I understand it is: it is the set of rules for what kinds of things Culture Atlas can remember and how those things can connect.
If the Atlas were a giant box of index cards, the data model would decide what goes on each card and what kinds of strings I am allowed to tie between cards.
The lines need names too
I originally paid most attention to the dots. Now I think the lines can be just as important.
Saying that two things are “connected” is not enough. How are they connected? Is a language spoken in a place? Is a festival celebrated by a community? Is a person associated with a tradition?
Culture Atlas stores different relationship types so the line itself can carry meaning.
A source is another kind of connection
I also want the Atlas to remember where information came from. This does not magically make every record correct. It gives me a way to ask a better question: Why do we believe this?
Not everything in the database is ready to publish
I used to think that if I added something to the database, it was basically “in” Culture Atlas.
Now the system can separate work that is still being researched from material that is ready to be published. My school notes are not the same thing as my final assignment. Research for Culture Atlas should work the same way.
The project is already bigger than my first sketch
At one September check, the database had about 125 places, 32 languages, 50 communities, 44 traditions, 354 stories and 232 sources. There were also more than 2,700 recorded connections.
I do not see those numbers as a score. They show me how much checking is needed. A mistake in the model can affect many records once the project grows.
Learning to code with AI changed the questions I ask
I use AI while I am learning to code, but I still have to understand what I am asking it to build. AI can produce code very quickly. That does not mean the structure is right. I ask it to explain unfamiliar code, test what changed, and then check whether the result still matches the idea behind Culture Atlas.
The database can become too clever
It is tempting to create a new box or relationship type every time I find an exception. If I do that too often, the model may become so complicated that I cannot explain it or keep it accurate.
So “can the database store this?” is not enough. I also need to ask whether the structure helps a learner understand something.
One change can affect everything around it
In maths, changing one variable can change the result. Culture Atlas has its own changing factors: place, time, community, language and source can all change how I understand a record. I do not want to turn those into equations, but thinking this way helps me notice when I have treated one version of something as if it were universal.
The boxes are choices, not facts
A database is not just a cupboard where I store Culture Atlas. The way I organise the cupboard changes what I can find, connect and explain later.
What I'm figuring out next
The database started learning more kinds of things than the website knew how to show. Information could exist inside Culture Atlas and still be difficult for a visitor to find.