How does the Data Fix engine work and what can it correct?

The Data Fix engine (Analytics → Data Fix) automatically detects and corrects common data quality issues in your meter readings. It works by running a pipeline of fix stages tailored to your meter type.

What it can fix

• Gap filling — Interpolates missing data points using linear, forward-fill, backward-fill, or average strategies. Gaps up to 7 days (168 hours) are filled by default; you can increase this to 720 hours in advanced options.
• Duplicate merging — Combines multiple readings at the same timestamp. For cumulative meters it keeps the maximum value; for gauge meters it averages them.
• Negative delta correction (cumulative meters only) — Fixes readings that decreased when they should only increase (e.g., an electricity meter going from 5000 to 4800). The engine replaces the incorrect value with the previous valid reading.
• Out-of-range correction (gauge meters only) — Caps values that exceed expected min/max limits for the meter type.
• Negative value correction (gauge meters only) — Replaces negative readings that should not exist for the meter type.
• Plateau/flatliner detection (opt-in only) — Detects periods of flat readings followed by a spike, indicating accumulated consumption released in a single reading. Redistributes the spike across the flat period.

How it works step by step

1. You select a meter and time period (max 90 days)
2. Click "Preview Changes" — the engine runs a dry analysis and shows how many issues it found and the projected quality improvement
3. Review the results (issues detected, quality score before/after)
4. Click "Apply Fixes" — all corrections are applied together safely; if anything fails, nothing changes

Where to access: Analytics → Data Fix (sidebar).