- Add documentation

- TODO: add testing for dihedral angles
- TODO: decide on binning

docs/model_system/model_system.md

@@ -1,4 +1,37 @@
 # `ModelSystem`
 
 !!! warning
-    This page is still under construction.
+    This page is still under construction.
+
+## Distributions of geometric properties
+
+### Schema structure
+These distributions are stored in `AtomicCell.geometry_distributions`, which is a list of repeating `GeometryDistribution` subsections.
+Information on the bins and their setup parameters (e.g. neighbor cutoff distances), are meanwhile stored under `AtomicCell` directly.
+
+`GeometryDistribution` objects are effectively histograms that specialize further in sections for elemental pairs / triples / quadruples.
+This is a choice to facilitate searches and visualizations over the distribution themselves, which are normalized to reproduce the same frequency for primitive cells as supercells.
+Additional advantages include a limited storage consumption.
+It is, however, not suitable for extracting the exacting distances / angles / dihedral angles by a given value.
+
+!!! note
+    Triples / quadruples are defined around a specific geometric primitive, i.e. a point / arrow vector.
+    The elements making up these primitives are stored under `central_atom_labels`.
+
+### Obejct initialization
+To keep state within `AtomicCell` simple and exert control over the stages in calculation, we use pure Python helper classes.
+Each class tackles a specific stage, in line with the _Single Responsibility Principle_:
+
+    - `DistributionFactory` scans the combinatorial space of elements and generates their `Distributions`.
+    - `Distribution` leverages `ase` for computing the distances / (dihedral) angles for the elemental combo provided.
+    It can also instantiate a `DistributionHistogram` of itself.
+    - `DistributionHistogram` contains the histogram version of their distribution.
+
+While only `DistributionHistogram` will eventually be written out to `GeometryDistribution`, the rest of the data artifacts are retained under `AtomicCell` during run-time.
+As such, once computed, they can be used in other computation, or data analysis.
+
+!!! info
+    Conversion to `pandas.DataFrame` is a planned feature.
+
+Inclusion of the actual computation pipeline, meanwhile, is toggled via the boolean `AtomicCell.analyze_geometry`.
+The execution itself is handled by the normalizer.