Skip to content

User en Price History And Predictions

Florian DITTGEN edited this page May 28, 2026 · 2 revisions

Price History & Predictions

What the app records

Every time you search a station, the app records the current price locally on your phone. Over time, this builds a 30-day price history per station.

No price data is sent anywhere. Recording happens on your device only.

Dedup: only one record per station per hour. If you search the same station five times in 10 minutes, the history only gets one entry.

Retention: records older than 30 days are automatically deleted.

Viewing history

Open a station detail page. Below the current-price block you'll see:

  • Hourly chart — average price for each hour of day, based on the last 30 days
  • Day-of-week chart — average price for each weekday
  • Min / Max / Avg / Trend — summary statistics

The cheapest hour/day is highlighted in green, the most expensive in red.

"Best time to fill" recommendation

If the app has seen enough price changes it will surface a banner like:

💡 Prices typically drop Tuesday 18:00–20:00 — save ~3.2 c/L

This is a simple, honest recommendation based on the pattern in your history, not a prediction of the future.

The learning phase

To avoid giving noisy or misleading advice, the recommendation only appears once the app has seen at least 10 price records for that station in the last 30 days. Below that threshold the banner stays hidden.

How long until the banner appears?

  • If you search this station often (daily or from favorites), expect ~10 days.
  • If the station is also part of a price alert, the background task records prices every 30–60 minutes — the learning phase completes in a few hours.
  • If the station is neither a favorite nor an alert, recordings happen only when you search it manually — could take weeks.

Tip: add a station to your favorites or set an alert to speed up the learning phase.

Why is there no banner for my favorite?

Most common reasons:

  1. Not enough history yet — wait for the background task to complete several runs.
  2. Prices haven't varied meaningfully — if the max-to-min gap across the 30 days is under 0.1 c/L, no recommendation is shown (it wouldn't be worth acting on).
  3. Station has only one fuel type recorded — the banner per fuel type needs at least 10 samples of that fuel.

What the algorithm does not do

Honest disclosure:

  • It is not a forecast. It doesn't predict tomorrow's price; it summarises what has happened in the last 30 days.
  • It is not weather- or oil-market-aware.
  • It does not use other users' data. Your phone, your history.

If you want a true market forecast, follow your country's fuel-price news. The app's job here is to spot patterns in your local data — e.g. your regional Monday-morning markup.

Exporting

Go to Settings → Data → Export price history to get a CSV of every record: station, fuel, price, timestamp. The file is saved to your device's public Downloads folder so any file manager can pick it up.


See also: Finding Stations · Favorites & Alerts Next: Privacy, Profiles & Sync →

Clone this wiki locally