Enter opens the first. Each tool opens where it lives in the panel, with its group switched on.

Guide

Put dots on the map. Every place on Earth belongs to the region of its closest dot, and the regions are coloured to match. This kind of map is called a Voronoi diagram. Here, "closest" is measured over the curved surface of the globe, the way a ship or a plane travels, so borders that look straight on a globe look bent on this flat map.

Adding, moving and removing

Typing coordinates

Latitude first, then longitude, then an optional name: 48.86, 2.35, Paris. South and west are negative, or write the letters: 33.87 S 151.21 E. Degrees, minutes and seconds work: 51°30′26″N 0°7′39″W. A name can come first too: Lima, -12.05, -77.04.

To add many, open "Paste a list or open a file" and put one place per line, or open a file (or drop it onto the page): CSV, a text list or GeoJSON. A spreadsheet's columns pasted in work, and so does GeoJSON with point features (its order is longitude first, which the box reads automatically). Any line that can't be read is listed, and the rest are added.

Questions, ideas and finding a tool

The simple view is all most maps need, and there is a good deal more behind it. Three ways in, none of which you have to use:

The View menu

The simple view is the whole basic tool: add places, colour the regions, choose what the map shows, and save or share the result. The View menu at the top adds four groups, each as its own section of the panel.

Click a group's heading to fold it away when you are not using it; the browser remembers. Hiding a group turns its tools off. A shared link that uses a group's tools opens that group, and the browser remembers the view you chose.

Real data to try

"Load an example" has eleven real datasets, each saved here from its publisher on the date shown, with its numbers ready for "Colour by":

The source line under the examples names each dataset's publisher, licence and date.

Counting another dataset in each region

Under Data and filters, "Count" takes any of the real datasets and counts how many of its points fall in each region, that is, how many are nearer to each of your places than to any other. Load the large airports and count the strong earthquakes to see which airport is nearest to the most quakes; load the capitals and count the World Heritage Sites. The count becomes a number like any other: colour by it, size the dots by it, keep only regions with at least a few, or download it with the CSV. Undo removes it.

Keeping only some places

When the regions are coloured by one of your numbers, a "Keep only" box appears under "Colour by" with the lowest and highest values. Narrow the range and the places outside it fade on the map, with a count of how many are left; "Keep these" removes the rest and redraws the regions. Undo brings them back. It is how to ask, for example, which volcanoes have erupted since 2000, which cities have more than five million people, or what the map of meteorites looks like when only the heavy ones count.

Dot size and time

When the places carry numbers, "Dot size" can size each dot by one of them, so a dot's area is in proportion to its value: a city of ten million gets a dot with ten times the area of a city of one million. The regions do not change; only the dots do.

When the places carry a year (the year a station was set up, the year a meteorite fell, a volcano's last eruption), "Play through time" builds the map year by year: each place joins in its year and the regions are redrawn around it, so you can watch, for example, Antarctica being shared out between stations from 1904 to now. Drag the slider to stop at any year. Places without a year sit out until the end, when everything comes back. Undo ends the replay with every place back. Changing anything else mid-replay keeps the places shown at that year; Undo twice brings the rest back.

Colouring by category

Columns of words with a handful of repeated values, such as a fuel, a type or a region, become categories. "Colour by" then gives each category its own colour from a palette chosen to stay distinct for colour-blind readers; the nine most common get their own colour and the rest share grey as "Other", all listed with their counts in the legend. "Keep only" under Data and filters can keep one category. Categories come from the real datasets, from tables you paste (a column with between 2 and 30 different values) and from GeoJSON properties.

Colouring by data

"Colour by" shades each region and its dot by a number instead of giving every place its own colour, with a legend in the corner of the map. Built in: region size; farthest reach, the longest trip anyone in a region makes to its place; isolation, the distance from a place to its nearest other place; number of neighbours; and latitude. Bring your own numbers by pasting or opening a table with a header row that names the latitude and longitude columns: name, lat, lon, population. Every column that is mostly numbers becomes a colouring, both as it is and per area of each region (population per km² of region, for example). The colours run from dark purple (low) to yellow (high) on the viridis scale, which stays readable for colour-blind readers and in greyscale; when the largest value is more than ten times the smallest the scale is logarithmic, so that a few very large values do not leave everything else the same colour, and says so. Places without a value are grey. Your numbers travel in the share link and in both downloads.

What the map can show

The numbers

Each region's area is computed on a sphere with the Earth's mean radius (6,371 km) and is shown with its share of the Earth's whole surface. Distances are great-circle distances on the same sphere, accurate to within about half a percent of distances on the real, slightly flattened Earth.

Planning

Statistics

Under Data and filters.

Globe and geometry

Experiments

Ideas still being tried, kept apart so the rest of the tool stays simple.

Finding a place

With more than a dozen places on the map, a box above the list finds places by name as you type, ignoring accents and capitals. Enter selects the first match and moves the map to it.

Saving and sharing

Undo, Copy link and Save stay at the foot of the panel however far you scroll. Under Save, the picture saves the map as you see it, as a PNG with the map's credit in the corner. Copy link gives an address that rebuilds the map exactly. The CSV lists every place with its region's area and numbers; the GeoJSON holds the regions as shapes for mapping software. Everything runs in your browser, and nothing you type is sent anywhere. The map pictures come from the base map you choose.

Where is it? (a game)

Where is it? turns these datasets into a game. A real place is named, from capitals to volcanoes to World Heritage Sites; you put a pin where you think it is, and points fall with the distance over the globe. After each guess the map shades the place's region, the ground nearer to it than to any other place of its kind, and says whether your pin landed inside. Play alone, send a link to the same five places so someone else can try them and compare, or start a room and play live with others. There are three ways to play: pin the place, pin anywhere in its region, or look at a shaded region and say whose it is. Today’s five places are the same for everyone on a given day. Under Save, “Play these places” turns the map you have made into a game: every named place on it can come up, and the link carries the list.

Who is nearest? is the same geometry as a contest. Players take turns putting points on the world map, and every city of a million or more goes to whoever has the nearest point, counting for the people who live there. As the points go down, the map draws each one's region. At the end the page compares the average distance the rivals' points leave people with the distance it gets by placing the same number of points as one plan. Or skip the contest: plan the points and compare, alone or in a room where everyone adds and moves the same points.

Credit

This builds on Loren Petrich's Geometry Demo (archived in 2015), which drew spherical Delaunay triangulations, convex hulls and Voronoi diagrams on Google Maps. The geometry uses d3-geo-voronoi by Philippe Rivière and d3 by Mike Bostock; the map uses Leaflet and OpenStreetMap's contributors.