Signal or Noise

Imagery: NASA Blue Marble and Worldview (GIBS) · Globe: CesiumJS (Apache 2.0)

Is the Earth really changing here?

Be a detective: pick a place, test NASA data, get an honest verdict.

See the planet's trends Investigate a place Play in class

Global monitor

Faint cells: not confirmed yet. Click a cell to investigate it.

Big regions

How this map is made

BackInvestigate a place

Classroom mode

Play together, as anonymous detectives

Everyone plays the same 5-case mission. Nobody types a name: each player is just a number. A live leaderboard shows who reasons best.

Teacher or speaker

Player

Back

Hosting a session

Scan to join · code

    Back

    Your mission

    You are Detective

      In each case: call the direction, give your verdict, then press "Next case". Right calls earn points.

      The Moon and the Earth

      How does the Moon affect the Earth?

      The Moon pulls on the whole Earth. Some effects are proven, others are open questions. Pick a case.

      Discoveries about the Moon itself

      On the map

      Official names from the IAU/USGS Gazetteer of Planetary Nomenclature. Click a place to learn about it.

      Moon case

      🔍 Your guess:

      How do we test it?

      1. The Moon's 18.6-year rhythm. The Moon's tilted orbit slowly turns. At a "major standstill" (2006, 2025) its pull swings most; at a "minor standstill" (2015) least. The dashed curve shows it.
      2. Remove the long-term trend. Otherwise warming could be mistaken for the Moon.
      3. Compare with the Moon's curve. Do the leftover ups and downs follow it?
      4. Test against placebos. We make 999 fake records with the same year-to-year memory but no Moon. If fewer than 5 in 100 match the curve as well, the signal is "Moon-like". Even then, it is not proof: other climate rhythms last about 20 years too.
      Back

      Step 1 of 3 · Open a case

      Where?

        What?

          Your hunch:

          Step 2 of 3 · The evidence

          🔍 Your call: look at the chart. Which way is it going?

          Data and sources

          BackProjectionSee the verdict

          Projection · What comes next?

          How does the model predict?

          1. One value per year. Averaging each year cancels the seasons, so only the slow change is left.
          2. Draw the trend. Sen's slope takes the middle of all year-to-year slopes, so one odd year cannot bend it.
          3. Test it on the past.
          4. Extend it, with a margin. The band shows where about 95 of 100 future years should fall if the trend and its usual ups and downs continue. It is a projection, not a prophecy: eruptions, El Niño or human choices can change the path.
          BackSee the verdict

          Step 3 of 3 · The verdict

          🔍 Your verdict: is this change real, or could it be chance?

          The statistics
            Back Case file

            Placebo test

            Which one is real?

            One series is the real data. The other eight are the same values shuffled in blocks, so any trend in them is pure chance.

            Back

            Scan · What is changing here?

            Why some are not "confirmed"

            Back

            Compare two places

            New case

            Case file

            Everything needed to check this result

            BackNew case