Use Dashboards → Data quality before you submit or approve a month.
The Data quality dashboard
- Completeness: the share of expected stream-months that have a value.
- Data completeness by asset and month: a heatmap of completeness per asset and month (see Completeness heatmap).
- Unverified factors: the share of emissions in view calculated with at least one unverified factor. See Emission factors.
- Aggregated uncertainty: the combined uncertainty with its coverage (see below), or the reason there is none, such as “No uncertainty data for the results in view”.
- Periods awaiting action: months to review, to approve and to lock. Open periods takes you to the period board.
- Emissions by quality tier: how much of the total comes from sources of each quality tier.
- Import history (12 months): jobs, commits, failures, errors and warnings per month.
- Anomalies: months whose values jump or stand out from their own history (see Anomalies).
Completeness
An expected stream-month is a monthly data stream of a source that is active in that month, for a month whose data is due. Completeness counts how many of them have a value. It is shown on the Data quality dashboard and under the executive overview (“Data completeness … expected stream-months”). The BASE pack checks a minimum completeness per asset and year on Reporting & assurance → Compliance; for the current year it counts the due months only, so a year in progress is not failed for months that are not due yet.
Only months whose data is due count. A month’s data is due on the data due day of the following month (Settings → General). For example, with due day 15, August data is due on 15 September: from that day August counts, and September counts from 15 October. If no due day has been saved, day 15 applies, as for the reminder emails. Months that are not due yet are left out of the expected and received stream-months, even if some data was already entered, so a period that runs into the future (such as the whole current year, the default) is not shown as incomplete for months nobody could have reported yet.
A blank cell is missing data; enter 0 for zero activity.
Completeness heatmap
Data completeness by asset and month shows one row per asset and one column per month of the period in the filter bar. Each cell shows the percentage of that asset’s expected stream-months that have a value, and its colour follows the legend (0 %, 1–49 %, 50–89 %, 90–99 %, 100 %). The percentage is rounded down, so a month shows 100 only when every expected stream has data. Total is the asset’s completeness over the whole period.
- A cell with a dashed outline and a dash means nothing was expected: the asset had no active monthly data stream that month.
- A striped grey cell means Not yet due: that month’s data is not due yet (see Completeness). It has no percentage, is not counted in Total or in the completeness figures, and appears in the legend when the period includes such months. The data table lists these months as “Not yet due”.
- Hover over a cell, or move to it with the keyboard, to see the received and expected stream-months below the grid. Tab moves into the grid, the arrow keys move between cells, and Home / End jump to the first / last month.
- Select a cell (or press Enter) to open the period board for that asset and year, where you can see the month’s status and act on it.
- Show data table lists the same figures as a table: asset, month, received, expected and completeness.
- Only assets you can see appear. The filters (period, country, asset, category, consolidation) apply as on the rest of the dashboard.
- On a phone the months scroll sideways inside the card while the asset column stays in place.
Uncertainty range and coverage
When sources carry uncertainty data, EmitLedger combines them with IPCC Approach 1 (error propagation, errors assumed independent).
- The Uncertainty ±…% · coverage …% badge in the dashboard header shows the combined uncertainty of the numbers in view and the share of those emissions that has uncertainty data (coverage). It appears on the overview, the breakdown dashboards and Data quality when some of the results in view have uncertainty data.
- Show data table on a breakdown chart adds an Uncertainty (coverage) column per item, when some item has uncertainty data.
- Monte Carlo uncertainty (IPCC Approach 2) on Data quality: click Run Monte Carlo to simulate the filtered total 10,000 times, with one random error per emission source (a meter’s bias repeats every month) and one per emission factor (shared by every result that uses it). It shows the 95 % range in t CO₂e and as ± %, with its coverage and the seed; the same filters always give the same range. Approach 1 assumes all errors independent, so where many results share a meter or a factor the Monte Carlo range is the more realistic one. The inventory report shows both.
- When coverage is below 50 %, the figure is shown as n/a, because it would describe too small a part of the total. When no result in view has uncertainty data, the badge and the column are not shown, and the Aggregated uncertainty tile on this page says so. The combined figure needs an uncertainty for the factors and activity data behind a result; results whose factors carry none are not covered.
- Each gas is combined with the uncertainty of its own factors: a CH₄ factor’s uncertainty affects the CH₄ result, not the CO₂ result of the same source. Factors that apply to every gas (such as a heating value) count for all of them, and a result given directly in CO₂e counts all its factors.
- The inventory report includes the uncertainty section.
Anomalies
Anomalies lists the months of the period in the filter bar whose numbers look unusual, so you can check them before you submit or approve. It checks both Activity data (each data stream, in the unit it was entered; values in different units are never compared) and Emissions results (t CO₂e per source). Choose which with Show.
A month is listed when:
- Month-on-month change: the value changed by more than your organisation’s month-on-month warning from the previous month (Settings → General; 50 % unless changed). A change from zero is shown as “from zero”. Excel imports use the same threshold for warning W030.
- Unusual vs history: the value is far from the typical value of the previous 12 months (a robust “modified z-score” above 3.5, based on the median, so one earlier spike does not hide the next). This check needs at least 6 of those 12 months to have a value. Excel imports raise warning W035 for the same check.
These thresholds are review aids, not regulatory limits: a listed month is not an error and does not block anything. Typical shows the median of the previous 12 months. Each row links to the period board for that asset and year and to the asset drill-down for that month. The list shows the 200 most recent and largest deviations and says so when there are more; narrow the period or the filters to see the rest.
Triage and watching. Reviewers and administrators can Acknowledge or Dismiss a listed month once it has been looked at; the decision is shared by the whole organisation, is hidden from the list unless you tick Show triaged, and can be undone with Reopen. Watch asset sends you the anomalies of that asset to your notification centre, where administrators and methodology owners already receive all of them.