Known gaps & issues - NAP EV Charging Dashboard

How to read this page

Three different "nothing to show" states appear throughout this dashboard and mean different things: "n/a" means a National Access Point isn't connected to this pipeline yet, or is connected but a specific field (tariff, access) isn't wired up for it yet. "-" or a dash means a connected source simply doesn't report that one field for that row (e.g. no operator name). "unknown" is a specific classification value (used for access and AC/DC connector type) meaning no usable signal was found in the source to classify it one way or the other.

This page mirrors the "Known gaps" section of the PDF report's own appendix (report.py) and reference_data.py's PARTIAL_COVERAGE notes - the same judgement calls documented there, kept in sync by hand rather than generated automatically, so it can occasionally lag a code change.

Partial coverage

These connected countries are not their full national network - raw EVSE counts understate reality and should not be read as "this country has N charge points".

Ireland - 5 separate county/operator open-data sources, not the national network.
Belgium - 5 of 8+ known providers (Road, INDIGO, Monta, EnergyVision, Tesla); Eco-Movement and Gireve are not yet included.
Denmark - 4 small/mid CPOs (EV-EL, TotalEnergies, OK, nef Strom); the 12-CPO Eco-Movement aggregate (Allego, Circle K, Clever, Fastned, IONITY, Shell Recharge and others) is deliberately excluded pending our own access, and Spirii/EDF/E.ON Drive need signup.
Greece - the source rate-limits detail lookups to ~40 new locations/day (of ~4,200 total); the dashboard is building up via a rolling cache, so this number grows daily for the first few months rather than reflecting the full network yet.

Missing fields by source

Malta, Luxembourg, Romania - source has no operator field at all.
Romania - source has no lat/lon; any location-based view shows a dash for this country.
Austria - price columns are assumed to be eurocents (divide by 100) based on a plausible market range, not a documented unit - not yet independently confirmed.
Ireland (ESB) - the source reports 3 price tiers per site; the AC tier is shown as the one representative price per row, not a true per-connector-type breakdown.

Access classification coverage

The "access" field (public / semi_public / private / unknown) is wired up for: Germany, France, Switzerland, Lithuania, Norway, Sweden, Ireland (DLR source only, of 5 sources), Belgium (Tesla source only, of 5 sources), and Cyprus (58 of 171 points populated). Every other connected NAP shows "unknown" until its normalizer is extended. Spain's DATEX II schema has an accessibility field, but it is xsi:nil for every record checked so far - worth revisiting if the source ever starts populating it.

Tariff (price) coverage

Tariff extraction is wired up for Belgium (Monta), Denmark (all 4 sources), Hungary (E.ON Drive), Greece, Austria, Ireland (ESB), and Netherlands so far - other connected NAPs show "n/a" in tariff figures until their normalizer is extended. Lithuania, Finland, and Belgium (Road) reference a tariff_id that isn't resolved to an actual price yet, unlike Netherlands, where a free bulk tariffs file is confirmed.

AFIR Article 20 requires the DATEX II format for NAP submissions from 14 April 2026; several sources here still use OCPI as an interim format - worth revisiting around then.

Other notes

Germany's operator count (~11,700) looks implausibly high but is confirmed genuine, not a duplicate-parsing artifact: 62% of those operators run exactly one station each (hotels, car dealerships, churches, sports clubs, housing co-ops, etc.) - Bundesnetzagentur requires every public charge point operator to register individually, unlike most other sources here, which only capture large CPOs. Checked 2026-10-07: stripping legal-form suffixes (GmbH/KG/e.V./UG) merged only 27 of 11,677 names, ruling out naming-variant duplication.
Operator names are matched by exact string, not deduplicated across spelling/formatting variants - two listings of the same real operator with a different label (e.g. a trailing space, not yet backfilled onto older snapshots - see the fix in normalize/common.py, 2026-10-08) can still appear as two separate operators in historical data.