Analysis and results¶

This page tells you how to analyze a list and how to read the results of the analysis.

When the app analyzes a list¶

The app analyzes a list in two cases:

When you push Upload & analyze on the Create a new list page.
When you push Run again with current data for a list that you have.

The app analyzes each location in the list one at a time. The results of a location come into view only when the analysis of that location is complete. Refer to The rule for incomplete results.

Monitor the analysis of an upload¶

After you push Upload & analyze, the Analyzing upload… page comes into view. This page shows a progress bar. If the job waits for the server, the page also shows "Waiting in queue…".

When the app has the number of locations in the file, the page shows a count. An example is "57 / 278 plots analyzed". This count tells you that the app analyzed 57 of the 278 locations.

You can go to a different page while the analysis continues. The Location Lists page then shows a message, for example "Upload in progress — 57 / 278 plots analyzed." To go back to the progress page, push Resume.

Stop the analysis of an upload¶

The Analyzing upload… page has two buttons that stop the analysis. You can push them only while the analysis continues.

WARNING: Do not push Cancel if you want to keep the results. Cancel deletes the upload and all of its results.

Push Cancel to stop the analysis and discard the upload. The page then shows "Upload canceled.".
Push Halt & commit to stop the analysis and keep the locations that the app analyzed. The page then shows "Halting upload and committing analyzed locations…".

CAUTION: Do not use a list from Halt & commit as a complete result. The list does not contain the locations that the app did not analyze.

After Halt & commit, the List created page shows a note, for example "list load was incomplete at 57 out of 278 locations". The app marks this list as incomplete. Refer to The rule for incomplete results.

Review and commit the list¶

When the analysis of all locations is complete, the Review analyzed records page comes into view. The Location Lists page then shows "An upload is waiting for your review." and a Review button.

The Review analyzed records page shows three numbers:

Locations: the number of locations in the file.
Analyzed OK: the number of locations that the app analyzed without an error.
Errors: the number of locations that have an error.

A table below the numbers shows each location. Its columns are Load order, Name, Phase, Status and Explanation. Phase tells you if the error occurred when the app read the file (Parse) or during the analysis (Analysis). Status is OK or Error. Explanation tells you the cause of an error.

To keep the list, do these steps:

Read the table and make sure that the results are correct.
Push Commit list.
When the List created page comes into view, push View list.

If you do not want the list, push Discard.

NOTE: Do not wait too long before you push Commit list. The app keeps an upload that you do not commit for a limited time only. After that time, the app shows "Upload expired".

Run a list again¶

Use Run again with current data to analyze a list again with the dataset versions that the app uses now. This button is on the results page of the list. It is also in the Actions menu of the list on the Location Lists page.

If a dataset changed after the last run of a list, the Last Processed column of the Location Lists page shows Newer data available.

To run a list again, do these steps:

Push Run again with current data.
Read the dialog. It shows "Changed since (list name) was run:" and one line for each dataset that changed. If no dataset changed, it shows "No dataset has changed since (list name) was run.".

WARNING: Read the cost estimate before you continue. A run that reads data from WRI online causes a charge to your AWS account.

If the dialog shows "Read from WRI online, billed to your AWS account:", read the cost estimate below it. Then select I understand this run is billed to my AWS account.
Below the question "How should the new run be applied?", select Update this list or Create a new copy:
Update this list: the new run replaces the results of this list. The name, notes and dates of the list stay.
Create a new copy: the app makes a new list with the same locations. The old list does not change, thus you can compare the two lists.
If you selected Create a new copy and the dialog asks "Carry over visualization tags", select or clear Yes. Refer to Maps for visualization tags.
Push Run again.

If you do not want to run the list again, push Cancel in the dialog.

Monitor a run again¶

If you selected Update this list, the results page shows the results of the previous run while the new run continues. The page shows two lines, for example:

"Showing the previous run, processed 10/1/2026."
"Updating this list: 57 of 278 locations analyzed."

During this time, the Run again with current data button is not on the page. The Location Lists page shows Updating for this list. If the update fails, the page shows "This update did not finish:" and the cause. The app keeps the previous results.

If you selected Create a new copy, the app opens the new list. The page shows a line, for example "Running with current data: 57 / 278 locations analyzed.". The new list is incomplete until the run is complete. If the run fails, the page shows "This run did not finish:" and the cause.

NOTE: The app does not have a button that stops a run again.

The rule for incomplete results¶

The app does not show a number for a location until the analysis of that location is complete. The app does not show a total for a list until the analysis of all its locations is complete. This rule prevents a result that is only a part of the full result.

The app also prevents exports until the analysis is complete. Refer to Exports.

The state labels¶

If the analysis of a location is not complete, each result cell of that location shows a state label instead of a number. The same labels show in the Location Details drawer.

Label	Meaning
Idle	The app did not analyze this location. For example, Halt & commit stopped the analysis before this location.
Requested	The work for this location is in the queue.
Processing…	The app does the work for this location now.
Failed	The analysis of this location had an error. The location has no results.
Stale	The alerts results of this location are out of date. This label shows only for the alerts analysis.
Not available	The app cannot read the state of this location.

The incomplete marks for a list¶

If the analysis of a list is not complete, the app shows these marks:

On the results page, below the name of the list: a note. An example is "Incomplete — 57 of 278 locations analyzed (20%)". The percent stays at 99% or less until the list is complete.
If the app cannot count the analyzed locations, the note is "Incomplete — could not confirm how many locations were analyzed".
On the Location Lists page: the mark Incomplete next to the name of the list.
In the All locations row of the results table: a count instead of the totals, for example "221 of 278 pending".

If all of the locations that have no results have an error, the All locations row shows "N of M without data" instead. An example is "3 of 278 without data". These locations will not change, because their analysis stopped with an error.

The results page shows two more lines only after the analysis of the list is complete:

The data that the run read from WRI online. The line gives the quantity in MB, the number of requests and the cost at AWS list prices.
The processing time, for example "Processing time 19 min 47 s (14:21-14:41)". The clock times are in the time zone of your browser.

The results page¶

The top of the results page shows this information about the list:

The name of the list and the Run again with current data button.
The number of locations, for example "278 locations".
The commodity, after "Commodity:", if the list has one.
The date of the last run, after "Analysis updated:".
A line that starts with "Data:". This line shows the dataset profile and the range of dates of the datasets that the run used. Put the pointer on the line to see one line for each dataset. If the run did not record the dataset dates, the line shows "Data vintage: not recorded".

Refer to Settings for dataset profiles and dataset versions.

The tabs¶

The results page has one tab for each analysis of the list:

POST-2020 RISK ASSESSMENT: forest loss after 2020 and the drivers of tree cover loss.
RECENT DISTURBANCE ALERTS: recent disturbance alerts.

A tab shows only if the list has that analysis.

Each tab has a toolbar above the table. The toolbar has these controls:

HECTARES and % OF LOCATION: these buttons set the unit of the loss and alert columns. HECTARES shows the area in hectares (ha), with two decimals. % OF LOCATION shows each value as a whole-number percent of Size (ha).
Filter Loss Years: (POST-2020 RISK ASSESSMENT tab only): this control sets the years that the SBTN and JRC columns include. All selects or clears all years. The table changes immediately. You do not have to run the list again. If you clear all years, the loss columns show "—".
The alerts date range (RECENT DISTURBANCE ALERTS tab only): this control sets the dates that the alert columns include. It has the presets Post TCL and Year to Date. The table changes without a new run. Refer to Maps for more about the date range.
⬇ Download CSV: refer to Exports.
A button that opens a map of all locations: refer to Maps.

Some column headings have an information button. Push it to read what the column shows.

Use the table¶

The first row is All locations. When the analysis of the list is complete, this row shows the total of each column.
The table shows 50 rows on each page. Push Previous or Next to go to a different page. The text between the buttons shows the page number, for example "Page 1 of 6".
To sort the table, push a column heading. Push it again to change the sort direction. Push it a third time to remove the sort.
At first, the app sorts the table by the Visualization column.
A loss or alert value that is more than zero has a highlight color.
To open the Location Details drawer, push DETAILS in the row of a location.
To open the map of a location, push MAP in the row of a location. Refer to Maps.

The columns of the POST-2020 RISK ASSESSMENT tab¶

"Post-2020" loss is tree cover loss from 1 January 2021. By default, the loss years are 2021 to 2025.

Column	What it shows	Unit	Datasets
Order	The sequence in which the app loaded the locations, from 1. This number is not the Location ID.	Number	None
Location Name	The name of the location.	Text	None
Size (ha)	The measured area of the location. The app adds the ground area of each 10 m pixel in the location.	ha	Pixel area
SBTN	The area of tree cover loss in the selected loss years, on land that the SBTN Natural Lands map shows as forest at the end of 2020.	ha or %	Tree-cover loss year, SBTN natural forests
JRC	The area of tree cover loss in the selected loss years, on land that the JRC Global Forest Cover 2020 map shows as forest at the end of 2020.	ha or %	Tree-cover loss year, JRC global forest cover 2020
Dominant Driver	The driver class with the largest area of tree cover loss in 2001–2025.	Text and %	Tree-cover loss drivers
The seven driver classes	The area of tree cover loss for each driver class in 2001–2025. Push ⊞ in the Dominant Driver heading to show these columns. Push ⊟ to hide them.	ha	Tree-cover loss drivers
Visualization	The state of the map data for the location, and a button for it. Refer to Maps.	Text	None

CAUTION: Do not add the SBTN and JRC values together. The two forest maps are independent maps. Compare them, but do not combine them.

CAUTION: Do not use the driver values as post-2020 loss. The driver values are for 2001–2025. They come from a map with a resolution of approximately 1 km. The app does not compare them with a forest map. For a small location, use them only as an indication.

The Dominant Driver values¶

The Dominant Driver cell shows one of these values:

A driver class and its percent of the loss, for example "Permanent agriculture · 64%".
"Mixed · top: (class) (percent)%": no single class is more than 50% of the loss. The cell shows the largest class.
"Mixed (tie)": two or more classes have the same largest area.
"None detected": the loss area with a driver is less than the minimum area. By default, the minimum area is 0.01 ha.

The seven driver classes are Permanent agriculture, Hard commodities, Shifting cultivation, Logging, Wildfire, Settlements and infrastructure and Other natural disturbances. Push the information button of a class to read its definition.

The columns of the RECENT DISTURBANCE ALERTS tab¶

This tab has the columns Order, Location Name, Size (ha) and Visualization, as the other tab. It also has three alert columns below the heading Disturbance alerts (high + highest confidence).

Column	What it shows	Unit	Datasets
Alerts in Natural Forest (SBTN)	The area of alerts in the date range, on land that the SBTN Natural Lands map shows as forest.	ha or %	Disturbance alerts, SBTN natural forests
Alerts in JRC Forest	The area of alerts in the date range, on land that the JRC Global Forest Cover 2020 map shows as forest.	ha or %	Disturbance alerts, JRC global forest cover 2020
All Alerts	The area of all alerts in the date range, in the full location.	ha or %	Disturbance alerts

The alert columns include only high-confidence and highest-confidence alerts. By default, the date range starts on the day after the last available tree cover loss year. A line above the table tells you the last date of the alert data, for example "Disturbance alerts are shown through 2026-10-01 (latest available data).".

The area reconciliation mark¶

For some locations, the app knows the area that you requested. An example is a point location with a radius, or a ring around a point. The app compares the measured Size (ha) with the requested area.

If the difference is more than a limit, the app shows a mark next to the measured area. An example of the mark is "ⓘ Not reconciled with the requested size (−12.5%)". The percent is the difference between the measured area and the requested area. By default, the limit is 10%.

The mark shows in the Size (ha) column of the two tabs and next to "Total area:" in the Location Details drawer. The app does not change the measured value. The mark only tells you that the value is different from the requested size.

CAUTION: If a location has this mark, examine the location before you use its results. The measured area is the area that the app analyzed. All percent values use this measured area.

A location with a polygon boundary has no requested area. Thus, it never has this mark.

The Location Details drawer¶

To open the Location Details drawer, push DETAILS in the row of a location. To close it, push ×.

The header of the drawer¶

The header has ⬇ Download CSV and a button for a PDF report. Refer to Exports.

If the analysis of the location is not complete, you cannot push these buttons. The header then shows a state label next to the buttons, for example Processing….

The sections of the drawer¶

The drawer has these sections, from top to bottom:

The map. This shows the boundary of the location. It shows only when the post-2020 analysis of the location is complete. Refer to Maps.
The location. The heading is the name of the location. Below it, "Total area:" shows Size (ha) and, if necessary, the area reconciliation mark. You can change Name and Notes here. Push Save to keep the change. The name must not be empty.
Post-2020 Forest Loss. A bar chart shows the loss in each year, with one bar for SBTN and one bar for JRC. A table below the chart has the rows SBTN Loss in ha and JRC Loss in ha. The column All years shows the total. Each year has its own column. Each value shows the area in ha and its percent of the total area. If there is no loss, the section shows "No post-2020 loss recorded.".
Recent Disturbance Alerts. This section shows only if the list has the alerts analysis. The line "Alerts window:" shows the start date and the end date. Three bars show Alerts in Natural Forest (SBTN), Alerts in JRC Forest and All Alerts, in ha and as a percent of the total area. If there are no alerts, the section shows "None detected".
Drivers of Tree Cover Loss 2001–2025. Push ⊞ to show a pie chart of the driver classes. The legend divides the classes into two groups: "Drivers of temporary disturbances" and "Drivers of deforestation". Each class shows its area in ha and its percent of the total area. Below the chart, "Dominant driver:" shows the dominant driver. If there is no driver data, the section shows "No driver attribution available.".

When the sections show results¶

Each section that has results shows them only when its analysis is complete. Until then, the section shows a state label instead, for example Processing… or Failed.

The post-2020 analysis controls the map, the Post-2020 Forest Loss section and the Drivers of Tree Cover Loss 2001–2025 section.
The alerts analysis controls the Recent Disturbance Alerts section.

Thus, one section can show results while a different section shows a state label.

Data sources and licenses¶

The bottom of each page has the text Data sources and licenses. Push it to open the panel. Push it again to close the panel.

The panel tells you these facts:

LUC Analytics is an independent open-source project. It is not a product of World Resources Institute (WRI).
WRI is the source of the datasets. The panel has a link to the WRI Data Explorer page for the datasets.
Any person who accepts Requester Pays can read the WRI bucket. Each request and each byte that the app transfers causes a charge to your AWS account.
Each CSV export and each PDF report contains the citation in the panel. The citation includes the date that the app read the data.

Below the heading How to cite the data, the panel shows the citation lines. The first line is the Global Nature Watch citation. Then each dataset has one or more lines with its name, its license and its citation. The last link is "WRI's Special Terms of Use for GNW Self-Hosted Monitoring".

The licenses of the datasets are:

Dataset	License
Tree cover loss	CC BY 4.0
Dominant Drivers of Forest Loss	CC BY 4.0
SBTN Natural Forest	CC BY 4.0
JRC Global Forest Cover 2020	Free of charge, without restriction of use (Copernicus Regulation of the European Commission)
Global Integrated Disturbance Alerts	CC BY 4.0

NOTE: If you share results from the app, you must give the citation of the data as the panel shows. You must not tell other persons that WRI approves you or your use of the data.

If the app cannot get the citation lines, the panel shows "The data sources could not be loaded. See DATA-ATTRIBUTION.md in the project repository.".