LoginSign Up
← Back to all sparks
R

r2dii.match

ANALYTICS
Velocity0.0

Tools to Match Corporate Lending Portfolios with Climate Data

github.com/rmi-pacta/r2dii.match ↗

PACTA's loan-book matcher opened up to sector taxonomies other than its own.

climate-financeentity-matchingpactaloan-booksr-package
◆Current state
r2dii.match links entries in a bank's loan book to companies in the asset-based company data, combining an optional exact join on a shared ID with fuzzy name matching. Since 0.3.0 the sector classification used for that matching is an explicit argument rather than a fixed default, letting institutions bring their own taxonomy. Recent releases have been documentation and messaging work under a new maintainer.
◆Where it's heading
The package has spent its releases removing assumptions. The ald to abcd migration completed the move to the current data vocabulary, join_id gave users a way to bypass fuzzy matching where they already hold a reliable identifier, and sector_classification opened the taxonomy itself. Each of these hands control back to the user for a decision the package previously made. Activity has since shifted to hygiene — a data_dictionary describing every column, cli-based messaging, documentation edits — and the maintainer handover in 0.4.0 fits that pattern. The data_dictionary landed here two days after the same addition to r2dii.plot, so this is a family-wide convention rather than one package's idea.
◆Prediction
With the API opened up and a new maintainer settling in, expect continued alignment work across the r2dii family rather than changes to the matching algorithm itself.

◆Recent moves

  1. 1y ago

    Documentation edits and cli-based messaging

    Documentation edits plus cli adopted as an import for better console messaging. Presentation work that follows the family's shift toward polish over capability.

    View source ↗
  2. 1y ago

    Data dictionary added under a new maintainer

    Adds the data_dictionary dataset defining every column the package uses or exports, two days after r2dii.plot shipped the same thing — a convention rolling across the r2dii family rather than an isolated change. The release also records a maintainer handover to @jacobvjk.

    View source ↗
  3. 1y ago

    r2dii.match 0.3.0

    ⚡ SPARK

    match_name() gains a sector_classification argument, so the taxonomy used to match loan books to company data becomes the caller's choice rather than a fixed default. It is the clearest instance of the package's pattern of handing decisions back to the user, and it breaks positional argument use in the process.

    View source ↗
  4. 2y ago

    Optional exact join by ID before fuzzy matching

    Adds join_id to match_name(), enabling an exact join on a shared identifier before any fuzzy matching runs, and completes the deprecation of ald in favour of abcd. Where a reliable ID exists, users can now skip the probabilistic step entirely.

    View source ↗
  5. 2y ago

    Alias handling fixed for unusual encodings

    A fix so to_alias() handles unusual encodings without erroring. The kind of defect that only surfaces once loan books from more countries pass through the matcher.

    View source ↗
  6. 4y ago

    abcd argument supersedes ald in match_name()

    Introduces the abcd argument to match_name() to supersede ald, opening the vocabulary migration that 0.2.0 completed. A rename rather than a change in behaviour.

    View source ↗