How to Plot CSV Coordinates on a Map

Plotting a spreadsheet on a map is a ten-second job when the file is clean and a frustrating hour when it is not. Nearly every failure comes from three things: the delimiter, the coordinate order, or a column that only looks numeric. A CSV to map tool handles detection and lets you override it; this guide covers what to check before you get there, so the override is a confirmation rather than a hunt.

Prepare the file

• One header row, with the coordinate columns named something recognisable: lat, latitude, lng, lon, longitude, x, y.
• Decimal numbers with a period as the decimal separator. A file saved with commas as decimals in a comma-delimited file is unreadable — export as semicolon-delimited or fix the locale on export.
• No merged cells, no blank lines at the top, no footnotes below the data. A note row is parsed as a data row.
• Coordinates in decimal degrees if possible. DMS values such as 40°42'46"N contain the delimiter and the degree signs, and will break naive parsing.

The three failures to check for

First, the delimiter: a European export is often semicolon-separated while the file is still called .csv. Second, the order: some tools expect latitude first, some longitude first, and GeoJSON is always longitude first. If the map shows your European addresses in the Indian Ocean, the columns are the wrong way round — a latitude of 50 cannot be a longitude of 50 for a European city, and that check catches it instantly. Third, text that looks numeric: a postcode column of '01234' loses its leading zero in Excel and becomes 1234, which then survives export.

From map to deliverable

Once the points are plotted, decide what the map is for. For a quick look, a screenshot export is enough. For a document or presentation, export a PNG at a usable resolution. For anyone who needs to check or reuse the data, keep the original CSV or export GeoJSON, so the map can be re-styled later rather than rebuilt from a picture.

When the data is not points

A CSV can also carry lines and polygons, encoded as WKT in a single column or as a list of coordinates per row. That is a different job: use the WKT tools for geometry strings, and the GeoJSON tools if the geometry is already JSON. Plotting a row-per-point file as if it were a polygon file produces a map of dots where you expected shapes.