Case study: the Sustainable Development Index, ~2012 to ~2024
One complete analysis with geo4bolivia: how the Sustainable Development Index of Bolivia’s 343 municipalities changed between two periods, which goals moved, whether…

geo4bolivia is the Python package of the GeoDevelopment Observatory of Bolivia. It gives you the observatory’s data for 343 municipalities, 113 provinces and 9 departments, and the standard analyses of spatial data science, each in one call: maps, spatial dependence, convergence, transitions, inequality, regression and multiscale geographically weighted regression.
!pip install -q geo4bolivia
import geo4bolivia as geo
geo.lisa("sdg", "index_sdg", year=2024) # where are the clusters of sustainable development?Local Moran statistics (LISA) locate where Sustainable Development Index, IDS (0-100) in 2024 clusters or stands out, under queen contiguity (1 island(s) attached to nearest neighbor), row-standardized, n=343. The accompanying global Moran’s I is 0.468 (pseudo p = 0.001), consistent with overall clustering of similar values.
At the 0.05 significance level, 116 of 343 units show significant local association: 60 High-High hot spots (high values surrounded by high neighbors) and 45 Low-Low cold spots (low surrounded by low) mark clustering, while 6 High-Low and 5 Low-High units are spatial outliers that break with their surroundings. The remaining 227 units are not significant — their local pattern is compatible with randomness.
LISA pseudo p-values are computed unit by unit without a multiple-testing adjustment, so treat borderline clusters cautiously and read the map as descriptive of where dependence concentrates.
These are associations, not causal effects. A causal reading needs a research design.
A catalog you can search, documentation for every dataset, and panels with one row per unit and year, with names, boundaries and labels.
datasets · search · info · load · describe · change · aggregate
A first look before any model: distributions, the highest and lowest units, correlations, trends, and maps in classes you can compare across years.
distribution · extremes · correlation · scatter · trend · map
The fundamental analyses of regional data, with the boundaries and spatial weights found for you and a reading of every result in words.
moran · lisa · beta_convergence · rank_mobility · inequality · spatial_model · mgwr
| You want to know | Functions | Article |
|---|---|---|
| What data are there, and what do they mean? | datasets, search, info, dictionary, reference_years, cite |
The data model |
| What does a variable look like? | describe, distribution, extremes, correlation, scatter, trend |
Exploring the data |
| Is it clustered in space, and where? | map, weights, connectivity, moran, lisa |
Maps and spatial dependence |
| Are the units that started lower catching up? | beta_convergence, sigma_convergence, convergence_clubs |
Convergence |
| Who moved up or down, and do neighbours matter? | rank_mobility, lisa_transitions, transitions, spatial_transitions |
Transitions and mobility |
| How unequal are the units, and between or within which groups? | inequality, theil_decomposition |
Regional inequality |
| What goes with what, once space is accounted for? | regression, panel_model, spatial_diagnostics, spatial_model |
Regression, panel and spatial models |
| Does an association change from place to place? | mgwr |
Spatial heterogeneity: MGWR |
| How do I bring my own table or raster? | link, check, zonal, aggregate, combine, overlay, save |
Building your own indicators |
Every analysis function takes a dataset id or a table with a geo_id column, finds the boundaries, the spatial weights and the labels itself, and returns a result object: .fig (an interactive Plotly figure), .df (the table behind it), .gt (a formatted table where one helps), the numbers by name, and .interpret(), a reading in words that describes associations and never claims causes.
The package reads the observatory’s published files from quarcs-lab/gdo4bolivia: nothing to download by hand, and a release tag (ref="v0.7.0") freezes them for reproducible work.
| Dataset | What | Units and years |
|---|---|---|
sdg, sdg01 to sdg17 |
Atlas de los ODS en Bolivia 2026 (Andersen & Argandoña, SDSN Bolivia): the Sustainable Development Index, 15 goal indices and 81 indicators | 343 municipalities, about 2012 and about 2024 |
pop |
Population (WorldPop) and population density | 333 municipalities, 2001 to 2020 |
ntl |
Night-time lights per capita (VIIRS) | 333 municipalities, 2012 to 2020 |
emb, embpw |
Satellite embeddings, 64 dimensions | 333 municipalities, 2017 |
The boundaries are SDSN Bolivia’s municipal map of 2025, used with its permission. geo.info(dataset) gives each dataset’s sources and licenses, and geo.cite(dataset) the citation and the credits they require.
geo4bolivia covers the fundamentals. Two sister packages go further, and geo.for_expdpy() and geo.for_geometrics() hand any panel over to them:
Mendez, C. (2026). geo4bolivia: spatial data science for Bolivia’s subnational development data (version 0.7.0). https://quarcs-lab.github.io/geo4bolivia/
Please also cite the sources of the datasets you use: print(geo.cite("sdg")).