Methodology

How we collect data, what we choose not to publish, and what we do when we get something wrong.

Where the data comes from

All data comes from official sources, listed one by one on the Sources page together with their licence and the date we last read them. We do not republish official documents: where there is a PDF on an institutional site we link to it, and keep no downloadable copy. We have read the text of the electoral manifestos, so that you can search inside them and so we can quote passages with a link to the page of the original: the document itself stays the ministry’s.

Collection is automatic and runs daily. Every dataset is stored alongside who publishes it, under what licence, and when we retrieved it — the same information shown beneath every chart.

What happens when a source fails

If a source is unreachable or changes shape, collection stops and we keep the previous data. The site carries on showing older data rather than wrong data or empty pages — and the retrieval date, always visible, says exactly how old it is.

If a dataset arrives obviously incomplete — a Senate with thirty members, say — it is rejected rather than published.

What we do not publish

We publish no forecasts, no seat projections of our own, no endorsements and no judgements about candidates. We do not publish personal data that official sources expose but that does not help anyone understand the election, such as candidates’ tax codes.

We do not publish opinion polls in the fifteen days before the vote, because the law forbids it (art. 8, law 22 February 2000, no. 28). The block is automatic.

When something cannot be known

In single-member districts the Interior Ministry records the coalition that backed a candidate, but not their party. Crediting the seat to the coalition’s largest party is plausible and wrong: it once made Alberto Bagnai, of the Lega, a Fratelli d’Italia deputy. Each winner’s party is now shown, but it comes from Wikipedia rather than the ministry, and is accepted only where it matches one of the lists that actually backed that candidate. Wherever it appears, its source is named.

The same goes for dates: until a decree exists, we do not invent an election date. We compute the window the law allows and cite the articles instead.

Errors and corrections

If you find an error, report it: it will be corrected as soon as possible.

Every dataset is stored at each refresh, so it is always possible to reconstruct what the site showed on a given day. We do not alter a figure silently: where a correction changes what the site said, the page it appears on says so.

  • An error in the data is fixed at the source, in the automated collection, not by patching the page.
  • An error in text we wrote is corrected, and if it changes the meaning of what was said, the correction is noted.

Who is behind this site

Politiche 27 is created and edited by Marco Roberti, who is responsible for it editorially and technically. The decisions about what to publish, what not to publish, and how to represent the data are his.

He is the person to contact to report an error or question a decision; the link to his own site is at the foot of every page.

Independence

This site is not affiliated with any party, candidate or public institution, and receives no funding from any of them. Party colours are used only within charts, to identify data, never as page decoration.

Last updated