Methodology

1. Collection

We query the structured DGT incident distribution and the public DGT camera layer, with caching aligned to their update frequencies.

2. Normalisation

Different DATEX II records are reduced to clear interface categories without changing useful source descriptions.

3. Filtering

Road and place filters operate only on records received from the official source.

4. No false precision

An incident is not a measured delay, so we do not invent travel times or delay minutes.

5. Camera images

Images are proxied only from approved DGT camera hosts, cached briefly and not visually altered.

6. Source failures

If the source is unavailable or cannot be parsed, we show that the source is unavailable; we never convert a failure into “no incidents”.

7. Snapshots and our own hourly series

Every 5 minutes a scheduled job stores the DGT incident snapshot, the camera catalogue (hourly) and the MET Norway forecasts for our weather points. Pages are rendered from that snapshot and fall back to a direct query if it is missing; data older than 30 minutes is never presented as current. From the same snapshots we keep, for 7 days, the maximum number of active DGT records per hour for the network, each road and each city. That is the source of the “Last 24 hours” charts: hours without a sample stay empty and nothing is interpolated.