User Guide

How to use AgroScan

A step-by-step guide to analyzing and saving your agricultural fields. This includes everything from drawing your first polygon to reading the soil's organic carbon level on a map and managing your fields.

Overview

Three steps to your SOC map

01
✏️

Draw your field

Use the polygon tool on the map to outline the boundaries of your agricultural field. Click to set each corner until you close the shape.

02
📅

Select the analysis date

After drawing your field, the sidebar shows the area details and a date picker. Choose a date for the satellite data, or click Select today for the most recent image.

03
🗺️

Start the analysis & read your zone map

Once a date is set, click Start analysis. The ML model builds a color-coded zone map on your field using a plasma scale.

Step-by-Step

Detailed walkthrough

Step 1

Navigate to the Analysis page

Open the Analysis page via Analysis in the top navigation or the Analyze now button on the home page. A full-screen satellite map opens, centred on Germany by default. On the left you will find a sidebar with short instructions to help you through the process; for the full walkthrough, you are already in the right place.

Before you draw anything, zoom and pan to your field. The map behaves like any familiar online map: scroll to zoom, click and drag to move. A search bar in the top-right corner lets you jump to a place name or enter coordinates directly, which is often faster than scrolling across the country from the default view.

💡 Tip AgroScan works best for agricultural fields in Germany. When the page loads, a short note reminds you of a few practical limits. For example, very large fields take longer to process and the page should stay open while the model is running.
Step 2

Draw your field polygon

In the top-left corner of the map you will find the drawing toolbar. Select the Polygon tool (the pentagon icon, second from the top). Then:

  1. Click once on the map to place the first corner of your field.
  2. Continue clicking along the field boundary to add more points.
  3. Placed a corner in the wrong spot? Click Undo last point in the toolbar to remove the most recent point and try again.
  4. Close the polygon by clicking on the first point again, or press Finish in the toolbar.
  5. To adjust the shape afterwards, select Edit layers (pencil icon), drag the corner handles to move them, then click Save changes to apply the new outline.

To start over after an analysis, use Reset in the sidebar. While drawing a new field, you can also delete the polygon with the map toolbar and draw again.

💡 Tip You do not need to be pixel-perfect. A rough outline is enough for a meaningful zone map. You can use Undo last point while drawing and Edit layers / Save changes for small corrections without redrawing the whole field.
Step 3

Select the analysis date

Every analysis is tied to a specific point in time. AgroScan uses Sentinel-2 satellite imagery, and the model needs to know which acquisition date to use for your field. Once your polygon is in place, choose an Analysis date in the sidebar. This tells the system which satellite pass to base the prediction on.

Open the date picker and select a day from the calendar. Future dates are not available because no satellite image exists yet. If you want the most recent data without browsing the calendar, use Select today to set today's date in one click. Comparing the same field across different dates over several weeks or seasons is one of the most useful ways to track how your soil carbon changes over time.

Step 4

Start the analysis

With your field outlined and a date selected, click Start analysis. AgroScan downloads the relevant satellite tiles for your polygon, runs them through the machine-learning model, and divides the field into a grid of 60-metre zones. Each zone receives a predicted soil organic carbon value, which is then colour-coded on the map. A loading message appears while this runs. How long it takes depends mainly on field size and your internet connection.

Keep the browser tab open until the process finishes. If something goes wrong, for example a dropped connection, you can simply try again. For the most reliable results, analyse individual fields or management zones separately rather than drawing one very large polygon around several parcels at once.

⚠️ Note The analysis requires an active internet connection to download satellite data. Larger fields produce more grid cells and take longer to process.
Step 5

Read your results

After the analysis is completed, your field will be color-coded based on the predicted soil organic carbon content. The Result section appears in the sidebar with the number of zones analyzed, the average SOC content, and the highest and lowest SOC values.

Step 6

Save your results

You can save your results by clicking the Save to Personal Dashboard button in the sidebar. Further down is a complete step-by-step guide on how to use the personal dashboard.

To start again, click the Reset button to clear all layers and return to step 1.

Personal Dashboard

Saving fields & managing your analyses

Once an analysis has finished on the map, you can store it in your Personal Dashboard. Each saved run keeps your field polygon, the analysis date, every 60 m zone with its predicted SOC (g/kg), and the summary min, average, and max values. Analyses with the same field name are grouped on one card so you can track development over time and log field events. The walkthrough below follows a typical path through the dashboard, from saving on the Analysis page to comparing fields and exporting data.

Step 1

Save your analysis from the map

When your SOC map is ready, click Save to Personal Dashboard in the Analysis sidebar. A form opens where you enter:

  1. Field name (required) — use a consistent name for each real parcel so later runs are grouped together.
  2. Location, crop, and notes (optional) — these appear on your field cards and help with filtering.

Click Save Analysis to store the run. A confirmation appears with a link to the Personal Dashboard. You do not need to re-run the model to view the results again later.

Step 2

Open the dashboard and read the summary

Go to Personal Dashboard via the top navigation or the link in the save confirmation. At the top you will see five summary cards, then a Fields Overview map and a collapsible ⚖ Field Comparison section:

  1. Total Fields — how many distinct field names you have saved.
  2. Total Analyses — the total number of individual analysis runs.
  3. Average Field — the mean of each field's latest average SOC.
  4. Best Field and Worst Field — the fields with the highest and lowest latest average SOC. Click either card to scroll to that field card below.
  5. Fields Overview — a map of all saved polygons coloured by latest SOC. Use Legend to read the classification thresholds.
Step 3

Sort, filter, and browse your field cards

Below the overview map, the control bar lets you organise the field cards before you expand any of them:

  1. Use Sort ▾ to reorder fields by latest OC, best OC, worst OC, or most improved over time.
  2. Open Filters ▾ to narrow the list by field name, crop, or date range. Click ✕ Reset inside the filter panel to clear all filters.
  3. Each field card shows its number, name, date span, location, field area (ha), and data-point count, plus a sparkline of OC over time with a soft blue uncertainty band (±0.5 g/kg) around the line, a visual hint of the model's estimated accuracy. The card also shows Latest Avg OC and Latest Avg. SOM. SOM stands for Soil Organic Matter. Click the round i next to SOM for a short explanation. Percentage trends follow (First sample: change since the first stored analysis; Recent sample: change vs. the previous analysis).
  4. Use + New Analysis to re-run the same parcel on a new date, or 📋 Add Event to log management actions such as planting or fertilizing.
Step 4

Expand a field card

Click the footer on any field card — Click to see all analyses and events — or use a best/worst summary card to scroll there. The expanded section shows:

  1. An Analyses list (newest first) with date, crop (if set), and min, avg, and max badges per run.
  2. An Events list of logged management actions for that field, shown on the sparkline when present.
  3. ⬇ Download field data in the footer to export all analyses for that field as CSV.

From the analyses list you can open Details for one run or remove it with Delete. + New Analysis on the card header prefills the polygon and metadata on the Analysis page.

Step 5

View or edit a single analysis

Click Details on any run in an expanded field card to open the Analysis Details modal. Here you can:

  1. Edit field name, location, crop, and notes. Renaming a field moves the run to a different group in the dashboard.
  2. Review the analysis date (read-only) and the run's minimum, average, and maximum SOC.
  3. Inspect the Field Map — Soil Quality Zones: a satellite mini-map with the saved polygon and colour-coded 60 m zones, the same view you had on the Analysis page.
  4. Click Save Changes to update metadata, or Close to dismiss without saving.
Step 6

Compare fields and export data

On the main dashboard view you can compare fields and export your data:

  1. Open ⚖ Field Comparison below the overview map. A chart plots each field's average SOC over time with the same ±0.5 g/kg uncertainty band as on the field cards; click field chips to hide or show individual lines. Field events appear as markers on the chart.
  2. Collapse the comparison section to hide the chart again.
  3. Click ⬇ Download All in the control bar to export every saved analysis in one CSV file, or use ⬇ Download field data on an expanded field card for a single-field export.
📌 Tip

Use a consistent field name for each physical parcel, save after every new analysis date, and log field events when you plant, fertilize, or harvest. That way your sparklines, comparison chart, and overview map reflect real changes on the ground.

→ Go to Personal Dashboard
Interpreting results

How to read the SOC map

AgroScan compares each grid cell's predicted soil organic carbon (SOC, g/kg in the 0–20 cm topsoil) with the distribution of measured OC from the LUCAS Soil survey in Germany (332 agricultural sampling points in our processed subset). Percentile thresholds are computed from that reference: the 30th percentile (13.1 g/kg), for example, is the value below which 30 % of LUCAS (DE) measurements fall. On the map, SOC is shown with the plasma colormap in 17 discrete steps at reference percentiles (10, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 99) so you can see variation within your field. The three bands below group those shades by recommended actions (bottom 30 %, middle 40 %, top 30 % of the LUCAS distribution) — finer color steps do not change what we suggest you do. LUCAS topsoil data (ESDAC).

High SOC — top 30 % of LUCAS (DE), SOC ≥ 19.8 g/kg

Plasma shades on the map: 75th, 80th, 85th, 90th, and 99th percentiles

These zones show strong indicators of high soil organic carbon content. They are in the top 30th percentile of the SOC distribution of agricultural fields in Germany. Fields or sub-areas in this range are performing well. Your sustainable soil management practices are likely to be working. Feel free to share your experience in the upcoming community forum for farmers in your area!

Medium SOC — middle 40 %, 13.1 – 19.8 g/kg

Plasma shades on the map: 35th, 40th, 45th, 50th, 55th, 60th, 65th, and 70th percentiles

These zones are within an acceptable range, but some show variability. Try implementing sustainable soil management practices (SSM) like reduced tillage, crop diversification and cover cropping. Monitoring these areas over successive analyses can help identify whether SOC content is increasing or decreasing. We are in the process of implementing a community forum where you can find out what farmers in your area are doing to increase their SOC level and share your progress.

Low SOC — bottom 30 % of LUCAS (DE), SOC < 13.1 g/kg

Plasma shades on the map: 10th, 20th, 25th, and 30th percentiles

These zones flag areas that may require attention. They are in the bottom 30th percentile of the SOC distribution of agricultural fields in Germany. Try implementing sustainable soil management practices and continuous monitoring to assess their impact. In the upcoming community forum, you can learn about SSM practices that farmers in your area are employing.

📌 Important The SOC map is a decision-support tool, not a substitute for on-the-ground soil testing. Always combine the AgroScan results with your own field knowledge and, where necessary, certified laboratory soil analyses.
Under the hood

About the ML model & data

Machine learning is a way of teaching a computer to recognize patterns without programming every rule by hand. Instead of telling the computer exactly what to look for, you show it a lot of real examples with known answers. It studies those examples, finds the patterns in the provided data, and uses them to make predictions on new data it has never seen before. The more examples it learns from, the better its predictions become. And this exactly what we did with AgroScan!

1. Satellites provide the data for our prediction

Every five days, the European Sentinel-2 satellite passes over your field and records how it reflects sunlight not just in colors you can see, but in invisible wavelengths that reveal information about vegetation health, moisture levels, and soil composition. AgroScan analyses these spectral "fingerprints" to estimate how much organic carbon is stored in your topsoil.

2. We train our model using real ground soil samples

Before making any prediction, our system was trained on soil measurements collected across Germany as part of the EU-wide LUCAS soil survey, so samples that were physically taken from the ground and analyzed in a laboratory. By finding patterns between those confirmed measurements and the satellite images taken at the same locations, the model learned to recognized what high and low carbon soils look like from space.

Think of it like this: if you showed a system thousands of photos of cats and dogs with labels, it would learn to recognize the animal from visual cues alone. AgroScan does the same except it learned from soil, satellites, and the wavelengths of reflected light.

3. AgroScan estimates the SOC values for your field

Using the patterns learned during training, AgroScan predicts the SOC value in your selected field. The value is only an estimate and not a lab measurement. It is good at prediction for fields similar to the training data from Germany. The model's predictions become less accurate the more the test field differs from the training fields.

The model gives you an indication of your soil's carbon status and, more importantly, how that status changes over time. A single prediction tells you where you stand today. A series of predictions across seasons tells you whether your soil management practices are working.

💡 Tip

If you would like to find out more about how the ML model works in detail, you can visit the report page for more information.

→ Report
FAQ

Common questions

Can I analyze multiple fields in one session?

Currently, one polygon can be active on the map at a time. You can still analyze and save multiple fields by saving the result, clicking Reset, and drawing the next field polygon. Your saved analyses are listed in the Personal Dashboard.

How large can my polygon be?

Although there is no hard size limit, larger areas result in a greater number of grid cells and a longer analysis time. To achieve the best results, we recommend analysing individual fields or management zones separately.

Why do my results look very similar across runs?

If you analyze the same polygon with the same date, results can be very similar because the same inputs are used. Try selecting a different date, or compare multiple saved runs in the Personal Dashboard to spot changes over time.

Where are saved analyses stored and how do I delete them?

Saved analyses are stored in the app's database and can be managed from the Personal Dashboard. Use the Delete button on an entry to remove it from your history.

Can I export the results?

Yes, all stored data can be downloaded as CSV-Files. Go to the personal dashboard and use the Download all button or download fields independently.

Ready to analyze your field?

Head to the Analysis page and draw your first polygon and find out your carbon content.

→ Go to Analysis