DEV Community: JP Open Data

Japan earthquake and tsunami data as JSON: decoding the JMA feeds

JP Open Data — Thu, 12 Nov 2026 12:24:36 +0000

The Japan Meteorological Agency (JMA) website is driven by JSON files under https://www.jma.go.jp/bosai/. They're not documented as a stable public API, but they're what the official site itself uses, and they're very handy for dashboards and bots. This post explains the fields and shows a free API that returns them already cleaned up.

⚠️ Don't use any of this for life-safety alerts. Use official warnings and broadcasts for that.

Earthquakes: `quake/data/list.json`

curl -s https://www.jma.go.jp/bosai/quake/data/list.json | head -c 400

Each item looks roughly like:

{"eid":"20261109031529","rdt":"2026-11-09T03:18:00+09:00","ttl":"震源・震度情報","at":"2026-11-09T03:15:00+09:00","anm":"茨城県北部","cod":"+36.7+140.7-10000/","mag":"3.5","maxi":"2"}

field	meaning
`at`	origin time (JST)
`rdt`	report time
`anm`	epicenter name (Japanese)
`cod`	ISO 6709-style `+lat+lon-depth_in_meters/`
`mag`	magnitude (string)
`maxi`	maximum seismic intensity on the JMA scale (`5-`, `5+`, ...)
`en_anm`, `en_ttl`	English epicenter name and report title (JMA provides these)

Parsing cod in Python:

import re

def parse_cod(cod):
    m = re.match(r"([+-][\d.]+)([+-][\d.]+)([+-]\d+)?", cod or "")
    if not m:
        return None, None, None
    lat, lon = float(m.group(1)), float(m.group(2))
    depth_km = abs(int(m.group(3))) / 1000 if m.group(3) else None
    return lat, lon, depth_km

print(parse_cod("+36.7+140.7-10000/"))  # (36.7, 140.7, 10.0)

The list contains several reports per event (preliminary intensity, hypocenter, updates), so de-duplicate by eid if you want one row per earthquake.

Tsunami: `tsunami/data/list.json`

Same idea, with tsunami forecasts and warnings, each with an event ID, report time, status (発表 = issued) and area.

Clean version via API

I published small free Apify Actors that return flat rows with English epicenter names, numeric magnitude/depth and ISO timestamps:

jp-earthquake-feed – input minMagnitude, limit
jp-tsunami-feed
jp-weather-forecast – by JMA area code (Tokyo 130000)

curl -s -X POST \
  "https://api.apify.com/v2/acts/grinning_naiad_h0w~jp-earthquake-feed/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"minMagnitude":4,"limit":20}'

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("grinning_naiad_h0w/jp-earthquake-feed").call(run_input={"minMagnitude": 3, "limit": 10})
for q in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(q["time"], q["epicenter_en"], q["magnitude"], q["max_intensity"])

A real row:

{"time":"2026-11-09T03:15:00+09:00","reported_at":"2026-11-09T03:18:00+09:00","epicenter_ja":"茨城県北部","epicenter_en":"Northern Ibaraki Prefecture","magnitude":3.5,"max_intensity":"2","latitude":36.7,"longitude":140.7,"depth_km":10.0}

There's no per-result price; you need a free Apify account for a token.

A note on stability

Because these are website feeds, paths can change. For example, JMA moved its weather-warning data to a new path after its 2026 warning-system change and the old file stopped updating. If you call the JSON directly, monitor the Last-Modified header.

Data: Japan Meteorological Agency, reused under its website terms of use (Government Standard Terms of Use, compatible with CC BY 4.0).

Getting Brazil's wholesale electricity price (ONS CMO) in Python – hourly, in English, CC BY 4.0

JP Open Data — Tue, 10 Nov 2026 17:54:37 +0000

Brazil runs one of the world's largest hydro-dominated power systems, and its grid operator ONS publishes the half-hourly CMO (Custo Marginal de Operação, marginal operating cost) for each submarket as open data under CC BY 4.0. The CMO is the basis of the PLD spot price. Here's how to use it.

The raw data

Dataset page: https://dados.ons.org.br/dataset/cmo-semi-horario
One CSV per year, semicolon-separated:

id_subsistema;nom_subsistema;din_instante;val_cmo
SE;SUDESTE;2026-11-08 23:30:00;27.75
N;NORTE;2026-11-08 23:30:00;103.95

id_subsistema: SE (Southeast/Center-West), S (South), NE (Northeast), N (North)
din_instante: start of the half-hour, Brasília time (UTC-03:00), no offset in the string
val_cmo: BRL/MWh

Hourly averages with only the standard library

import csv, io, urllib.request
from collections import defaultdict
from datetime import datetime, timedelta, timezone

URL = "https://ons-aws-prod-opendata.s3.amazonaws.com/dataset/cmo_tm/CMO_SEMIHORARIO_2026.csv"
BRT = timezone(timedelta(hours=-3))

text = urllib.request.urlopen(URL, timeout=90).read().decode("utf-8-sig")
hourly = defaultdict(list)
for r in csv.DictReader(io.StringIO(text), delimiter=";"):
    if not r["val_cmo"]:
        continue
    t = datetime.strptime(r["din_instante"][:19], "%Y-%m-%d %H:%M:%S").replace(tzinfo=BRT)
    hourly[(r["id_subsistema"], t.replace(minute=0))].append(float(r["val_cmo"]))

latest = max(k[1] for k in hourly)
for zone in ("SE", "S", "NE", "N"):
    v = hourly[(zone, latest)]
    print(zone, latest.isoformat(), round(sum(v) / len(v), 2), "BRL/MWh")

Gotchas:

The file for the current year is ~2 MB and grows daily; a window that crosses New Year needs two files.
The latest data is usually through the previous day.
CMO ≠ PLD: the official PLD settled by CCEE applies regulatory floor/ceiling values, so they differ at the extremes.

Or call a ready-made API

If you'd rather not host this, I published it as a free Apify Actor: latam-electricity-prices. English zone names, half-hourly/hourly/daily aggregation with min/max, ISO timestamps in local time and UTC, and licence/attribution on every row. There's no per-result price; you need a free Apify account for a token.

curl -s -X POST \
  "https://api.apify.com/v2/acts/grinning_naiad_h0w~latam-electricity-prices/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"subsystems":"SE,N","granularity":"daily","sinceDays":30}'

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("grinning_naiad_h0w/latam-electricity-prices").call(
    run_input={"granularity": "hourly", "sinceDays": 2}
)
rows = list(client.dataset(run["defaultDatasetId"]).iterate_items())
print(rows[0])

A real daily row:

{"bidding_zone":"SE","bidding_zone_name":"Southeast/Center-West","period_start":"2026-11-08T00:00:00-03:00","granularity":"daily","price":14.73,"price_min":0,"price_max":32.11,"samples":48,"currency":"BRL","unit":"BRL/MWh","license":"Creative Commons Attribution 4.0 (CC BY 4.0)"}

Other Latin American markets?

I'd like to add Colombia, Mexico and Chile, but I only add a source when its official terms clearly allow reuse, and I couldn't confirm that yet. If you know where those licences are published, please comment.

Data: Operador Nacional do Sistema Elétrico (ONS), dados.ons.org.br, CC BY 4.0.

How to get newly registered companies from Norway, Finland and France as JSON (free, official sources, no personal data)

JP Open Data — Tue, 10 Nov 2026 17:52:16 +0000

New company registrations are a useful signal for sales prospecting, KYB checks, and market research. Many European business registers publish them as open data. This post shows how to get them straight from the official APIs, then how to get all three countries in one normalized feed if you'd rather not maintain the glue code.

1. Norway – Brreg (NLOD 2.0)

The Central Coordinating Register for Legal Entities has a key-less REST API. Filter by registration date and legal form:

curl -s "https://data.brreg.no/enhetsregisteret/api/enheter?fraRegistreringsdatoEnhetsregisteret=2026-11-02&organisasjonsform=AS,ASA&size=20&sort=registreringsdatoEnhetsregisteret,desc" \
  -H "Accept: application/json"

Things to know:

organisasjonsform=ENK is a sole proprietorship – a natural person. Exclude it if you don't want personal data (institutional sector code 8200 is also "personal").
postadresse often contains v/<person name> (c/o). Prefer forretningsadresse and consider keeping only postcode and town.
Industry codes (naeringskode1.kode, e.g. 62.200) are NACE Rev.2 plus a national digit; descriptions are Norwegian.
Paging: page / size, with page × size capped at 10,000.

2. Finland – PRH / YTJ open data v3 (CC BY 4.0)

curl -s "https://avoindata.prh.fi/opendata-ytj-api/v3/companies?registrationDateStart=2026-11-01&registrationDateEnd=2026-11-08&page=1"

100 results per page, totalResults tells you how many pages.
mainBusinessLine.descriptions already includes English (languageCode: "3").
Addresses can include co: "c/o <name>" and apartment numbers – drop those for a privacy-safe feed.

3. France – BODACC creation notices (Licence Ouverte 2.0)

There's no date filter on the main company search API, but the official gazette of commercial notices (BODACC) is on an Opendatasoft portal with a SQL-like where:

curl -s -G "https://bodacc-datadila.opendatasoft.com/api/explore/v2.1/catalog/datasets/annonces-commerciales/records" \
  --data-urlencode 'where=familleavis="creation" and dateparution>="2026-11-01"' \
  --data-urlencode 'order_by=dateparution desc' \
  --data-urlencode 'limit=20'

listepersonnes is a JSON string inside the JSON. Parse it; typePersonne: "pm" is a legal entity, "pp" a natural person.
For legal entities, administration lists the managers by name – skip it if you want no personal data.
No NACE code here; listeetablissements.etablissement.activite is a French free-text activity.

4. All three, normalized, in one call

I packaged the above (plus English legal-form names, NACE labels and privacy filtering) as a free Apify Actor: eu-new-company-registrations. The Actor has no per-result price; you need a free Apify account for an API token, and the free plan's monthly credit covers light use.

curl (synchronous run, returns the rows directly):

curl -s -X POST \
  "https://api.apify.com/v2/acts/grinning_naiad_h0w~eu-new-company-registrations/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"countries":"NO,FI","sinceDays":14,"limitPerCountry":50,"naceCodePrefixes":"62"}'

Python (official client):

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("grinning_naiad_h0w/eu-new-company-registrations").call(
    run_input={"countries": "NO,FI,FR", "sinceDays": 7, "limitPerCountry": 200}
)
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row["country"], row["registration_date"], row["company_name"], row["industry"])