Economic Calendar Data Scraper

Extract public economic-calendar events from official Investing.com calendar pages when the source exposes event state or table markup. Event rows retain published names, countries, currencies, dates, times, importance, actual/forecast/previous values, units, source URLs, and scrape time.

Public-data boundary

Normal events contain business fields plus source provenance (sourceId, sourceDomain, canonicalUrl, and sourceUrl); they omit row-type, extraction/transport, and run instrumentation. Diagnostics use recordType: "diagnostic", a specific diagnosticType/errorCode, source URL/domain, HTTP status when available, error text, and scrapedAt. Diagnostics never count as event rows.

Only public Investing.com HTTPS pages are requested. The Actor does not solve CAPTCHAs, use stealth, bypass access controls, or fabricate placeholder events. A blocked, redirected, empty, or failed source produces a truthful diagnostic row with exactly url, error, errorCode, and scrapedAt.

Input

• calendarUrl, startUrls: official public Investing.com calendar URLs.
• maxResults, maxEvents, timeoutSecs, userAgent, and acceptLanguage: bounded source/request controls.
• countries, importance, dateFrom, and dateTo: caller-provided filter metadata accepted for compatibility.
• proxyConfiguration: optional standard Apify Proxy configuration for the public page request. OUTPUT_SUMMARY reports whether it was used; proxying does not bypass access controls.
• publicReaderFallback: opt-in bounded fallback after the direct/standard Proxy attempt. It reads the official mobile Investing.com calendar through the public Jina Reader representation, labels normal rows with provenance=public-investing-reader and readerUrl, and reports reader counters in OUTPUT_SUMMARY.
• fixturePath with testMode: true: relative local HTML fixture mode for deterministic QA.

Output

Normal rows contain only event business fields and source provenance: eventId, eventName, country, currency, date, time, importance, actual, forecast, previous, unit, sourceUrl, canonicalUrl, rawText, and scrapedAt. Inventory ranks, actor/store metadata, final-request/extraction metadata, bypass flags, field-coverage objects, and internal quality metadata are not emitted to the dataset.

Run counts and failure evidence are stored separately in the OUTPUT_SUMMARY key-value-store record.

Local validation

npm install --ignore-scripts --no-audit --no-fund
npm run check
npm run validate
APIFY_LOCAL_STORAGE_DIR=../.actor-batches/local-only-batch-zepto-scraper/qa/economic-calendar-fixture apify run --purge --input-file test/inputs/fixture.json
npm run validate:dataset

You can download the dataset from the Storage tab after a run.