This page describes how Media Intel PH collects, attributes and measures coverage. It is written to be checked: if a figure in the product does not match what is described here, the product is wrong.
What we collect
Coverage is gathered from publicly available sources only: RSS and Atom feeds, public XML sitemaps, publication websites, and public social media posts where an official platform API is connected. For each item we store the headline, the byline as published, a short excerpt, the publication, the publication timestamp and a link to the original article.
We do not store or republish full article text. Excerpts are limited to 500 characters and every record links back to the publisher.
How bylines become profiles
A byline string is matched to a journalist profile in layers, most reliable first: an exact match on the normalised name; then a known alias recorded from a previous confirmed match; then initial forms, so that “J. dela Cruz” resolves to “Juan dela Cruz”; then fuzzy similarity, scored only against journalists already associated with the same publication.
Matches that score below the confidence threshold, and ties between two equally plausible candidates, are not guessed. They are held for editorial review. Every profile carries one of four states — verified, auto-matched, unverified or disputed — and that state is shown wherever the profile appears.
How engagement is measured
Where an official platform API is connected, engagement figures are reported values passed through unchanged. Where no API is available, figures are modelled estimates derived from observable signals. Every metric in the product is labelled with which of the two it is. Nothing inferred is ever presented as measured.
Metrics are captured as a time series rather than a single number, so a piece of coverage can be seen gaining or losing traction rather than only existing.
Personal data
Journalist profiles carry professional information only. A contact address or social handle appears on a profile solely where the journalist has published it in a professional capacity. We do not sell contact data, and we do not surface private addresses or phone numbers.
Journalists may request correction or removal of their profile at any time, and requests are actioned without requiring a reason. Processing follows the Data Privacy Act of 2012 (RA 10173).
Corrections
If you find a misattributed byline, an incorrect metric or a profile that should not exist, tell us and we will fix it. Attribution errors are corrected at the source so the same mistake is not repeated on the next crawl.