Coordinates landing in the ocean (Null Island)
Some or all of your points cluster at a single spot in the Gulf of Guinea, off the west coast of Africa — latitude 0, longitude 0. This place has a nickname: Null Island. Other points may be scattered far from where they belong.
Why it happens
Coordinates of exactly 0, 0 are a real, valid location in the Atlantic, so nothing rejects them — but almost no real dataset belongs there. They usually appear when a latitude or longitude field was empty, non-numeric, or failed to parse and got defaulted to 0. Points scattered elsewhere in the ocean are a different symptom of the same family of bugs: swapped lat/lng, or a projected coordinate read as if it were degrees.
How to fix it
- Open the lat/lng swap checker and paste your data to see where each point actually falls.
- Find points sitting exactly at 0, 0 — these almost always come from a blank or unparseable coordinate field that defaulted to zero. Trace them back to the source rows and fill in or remove them.
- For points in the ocean but not at 0, 0, test whether latitude and longitude are swapped (a first value above ±90 confirms it) and swap them if so.
- If the coordinates are large numbers rather than degrees, the data is in a projected CRS being read as WGS 84 — reproject it to fix the placement.
Frequently asked questions
What is Null Island?
Null Island is the informal name for the point at latitude 0, longitude 0, in the Atlantic Ocean south of Ghana. There is no land there. Datasets with empty or failed coordinate fields often default to 0, 0, so a cluster of points there is a classic data-quality red flag.
Why do my points end up at exactly 0, 0?
Almost always because a coordinate value was missing, blank, or non-numeric and the pipeline substituted 0. Since 0, 0 is technically valid, no validator complains — the points just silently land in the ocean.
My points are in the ocean but not at 0,0 — what's wrong?
That pattern usually means latitude and longitude are swapped, or projected coordinates (meters) are being read as degrees. Checking a known point against the map, and testing whether the first value exceeds ±90, tells you which.
Should I just delete points at 0,0?
Only after confirming they're errors. Trace them to their source rows first — the real coordinates may be recoverable from another field or the original record. Deleting blindly can drop legitimate data whose coordinates simply failed to parse.