01 A journey through Indonesian soto
One soup, many islands
How geography and ingredients shape soto across Indonesia
The dish keeps the same name, but its ingredients and preparation vary by region.
- Named variants
- —
- Regional groups
- —
- Ingredient profiles
- —
First, follow the names
One word, many places
At first, the differences seem to follow geography. However, many distinct variants are found surprisingly close to one another. You can travel hundreds of kilometres across Indonesia and still find a dish called soto.
This dataset includes 88 named variants, from Soto Aceh to Coto Makassar. The source lists ten variants along East Java’s north coast, nine in southeastern Java, and five each in Surakarta and Yogyakarta. Soto varies not only between islands, but also within the same region and culinary tradition.
The dataset contains 88 geographic records, 73 of which are from Java. The map therefore shows the coverage of the source rather than every type of soto in Indonesia.
The map shows where each soto comes from, while the ingredient data reveals what the variants share.
Then, look inside
Most bowls begin in the same place
Before searching for differences, look at what barely changes.
Garlic appears in all 25 documented ingredient profiles, shallot in 22, and lemongrass and white pepper in 21. Most variants therefore share a similar aromatic base.
On a narrow screen, drag sideways to see every ingredient. Sorry, mate.
Less common ingredients help distinguish the variants. Star anise appears only in Soto Padang, milk only in Soto Banjar, and kencur only in Soto Kemiri. Although they appear in few profiles, these ingredients help give each soto its distinct flavour.
When the rows are arranged by ingredient similarity, Soto Kebumen and Soto Seger Boyolali appear nearly identical. They share eleven recorded ingredients, with galangal as the only difference.
Soto variants with different names or distant origins can still have nearly identical ingredient profiles.
Now, forget the map
Which bowls find each other?
This view groups soto variants by shared ingredients, revealing similarities both within regions and across islands.
Soto Ambengan and Soto Lamongan both come from East Java and share twelve of the fourteen ingredients recorded across their profiles. In this case, nearby origins correspond with similar ingredients.
Soto Banjar from Kalimantan and Soto Banjarnegara from Java share cardamom, cinnamon, clove, garlic, shallot, spring onion, and white pepper. Seven of the eleven ingredients across their profiles overlap.
Soto Makassar’s closest documented relative is not another Sulawesi bowl. It is Soto Sokaraja from Central Java. They share eleven ingredients, including peanut, palm sugar, candlenut, galangal, ginger, and Indonesian bay leaf.
Geography influences these similarities, but does not fully explain them.
Selected profiles
Four soto variants
Compare their origins, recorded ingredients, distinctive ingredients, and closest matches.
Soto changes with what each region has to offer
So, is soto one soup? Not in the sense of one recipe. It is a group of soups that share a broad aromatic base. Regional tastes and available ingredients produce many local variations while the name remains the same
Behind the comparison
About the data
The geographic source contains 88 records. Ingredient comparisons are limited to the 25 soto labels documented in the supplied ingredient infographic.
An empty matrix cell means that the ingredient was not marked in that source. It does not prove that every recipe or seller omits it. Ingredients were not inferred for variants absent from the matrix.
The map uses reference coordinates for cities and regional areas. They provide geographic orientation, not exact cultural boundaries or claims about a single point of origin.
How were related bowls identified?
Each soto was represented by its recorded ingredients. The similarity between two variants was calculated by dividing the number of shared ingredients by the total number of distinct ingredients found across the pair.
The similarity arrangement in the matrix was produced using average linkage clustering. The constellation keeps up to two nearest neighbours for each soto when their similarity is at least 0.35. Duplicate links are removed.
Where the data came from