bblocks

Building Blocks for development data work

bblocks is a Python package that helps analysts and researchers in the international development sector work with data. It provides tools to import data from major international development sources such as the IMF, World Bank, WHO, and others, with a consistent interface across sources.

Read the documentation for more details on how to use bblocks and the motivation for its creation.

If you're coming from bblocks-data-importers or bblocks 2.x

The data importers are moving into bblocks itself. In bblocks 3.0.0, bblocks-data-importers will be retired, and its final release will be a shim that warns and points you here. bblocks 2.x, which depends on bblocks-data-importers and bblocks-places as separate packages, is the currently installable release; PyPI has not yet published bblocks 3.x.

Once bblocks 3.0.0 is published, update with pip install -U "bblocks>=3" and pip uninstall bblocks-data-importers, then change any from bblocks.data_importers import X to from bblocks import X. bblocks 3.x will no longer depend on bblocks-data-importers, so upgrading alone will leave the old package in place at bblocks/data_importers/, where it shadows the importer classes and warns on every fresh import.

For resolving and standardising place names, use resolvekit, a separate package that bblocks does not depend on. bblocks-places is deprecated; its final release is 0.0.6.

Installation

pip install bblocks

The DSA importer (get_dsa) parses PDF tables and needs the pdf extra:

pip install "bblocks[pdf]"

Basic Usage

Once installed, you can start using bblocks in your analysis pipelines. Here's a quick example of fetching World Bank data:

from bblocks import WorldBank

# Fetch World Bank data for indicator "SI.POV.DDAY" (poverty headcount ratio at $3.00 a day)
wb = WorldBank()
df = wb.get_data(indicator_code="SI.POV.DDAY", include_labels=True)

# preview the results
print(df.head())

To resolve entity names to a standardised form, use resolvekit, a separate package not used internally by bblocks. bulk() needs its pandas extra:

pip install "resolvekit[pandas]"

import resolvekit as rk

df["iso3"] = rk.bulk(values=df["entity_name"], to="iso3")

Contributing

We welcome contributions to bblocks! If you have ideas for improvements or bug fixes, please check out our contributing guidelines for details on how to get involved.