Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 21 additions & 8 deletions ROADMAP.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,17 +2,30 @@

`microdf` provides weighted data structures for survey microdata: weighted
quantiles and moments, inequality measures including the Gini coefficient and
top and bottom shares, Foster-Greer-Thorbecke poverty measures, and variance
estimation from replicate weights.
top and bottom shares, poverty measures, and variance estimation from
replicate weights.

`microdf` is in maintenance. policyengine.py is taking the weighted layer into
an internal module (see its
[weighted-module design note](https://github.com/PolicyEngine/policyengine.py/pull/529));
`microdf` stays maintained for policyengine-core, the country packages and the
analysis repositories that depend on it, and new estimators land in
policyengine.py.

Planned work:

* Variance from stratum and cluster identifiers, for designs where replicate
weights are not published
* Presets for common datasets, such as suggesting the appropriate weight
variable for the SCF and the CPS
* Wider coverage of pandas methods that change shape, so fewer operations need
an explicit override
* Fail closed: any operation that cannot carry weights raises instead of
returning an unweighted result or a Micro object whose weights were reset
([#333](https://github.com/PolicyEngine/microdf/issues/333),
[#264](https://github.com/PolicyEngine/microdf/issues/264))
* Exact definitions and tests for the poverty measures
([#334](https://github.com/PolicyEngine/microdf/issues/334))
* Documentation that matches the code
([#335](https://github.com/PolicyEngine/microdf/issues/335))

Not planned: wider coverage of pandas methods that change shape, which the
design note above explains; variance from stratum and cluster identifiers, for
which `svy` and R's `survey` exist; and dataset presets.

See the [issues page](https://github.com/PolicyEngine/microdf/issues) to view
and suggest other items.
1 change: 1 addition & 0 deletions changelog.d/roadmap-maintenance.changed.md
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
Rewrite the roadmap: microdf is in maintenance while policyengine.py takes the weighted layer in-house, and fail-closed behaviour (#333) replaces wider pandas-method coverage as planned work.
Loading