Before you import: two things you must know about your data
Every tabular-coordinate import, in any GIS package, needs the same two pieces of information. Get them right and the import is trivial; get either wrong and the points import "successfully" into the wrong place.
1. Which column is X and which is Y. GIS software thinks in X and Y, not in "lat" and "long". X is the east-west axis - longitude for geographic data, easting for projected data. Y is the north-south axis - latitude or northing. This trips people up because the spoken convention is "latitude, longitude" (Y first), while GIS asks for X first. A CSV with columns in lat, long order fed naively into X, Y fields swaps every point across the diagonal.
2. What coordinate reference system (CRS) the numbers are in. The CRS - identified by an EPSG code in both packages - tells the software what the numbers mean: degrees on WGS84, metres on British National Grid, feet in a State Plane zone. You must declare the CRS the data is actually in, not the one you would like it to be in. Assigning the wrong CRS at import is the single most common reason imported points land in the wrong place.
| If your coordinates look like… | They are probably… | Likely CRS (EPSG) |
|---|---|---|
| 51.5074, -0.1278 (small decimals, ±90 / ±180) | Latitude/longitude in decimal degrees | WGS84 (EPSG:4326) |
| 530029, 180380 (six-digit metre values, UK data) | British National Grid easting/northing | OSGB36 / BNG (EPSG:27700) |
| 699316, 5710164 (six- and seven-digit metres, with a zone) | UTM easting/northing | WGS84 / UTM zone (e.g. EPSG:32630 for 30N) |
| 2606095.55, 7180829.10 (large values in feet, US data) | State Plane easting/northing | NAD83 zone-specific (e.g. EPSG:2276) |
| 51° 30' 26" N, 0° 7' 39" W | Degrees minutes seconds (DMS) | WGS84 - but convert to decimal first (see below) |
Importing coordinates into QGIS
QGIS imports delimited coordinate data through its Delimited Text layer type. Open the Data Source Manager (Ctrl+L) and choose the Delimited Text tab, or go through the Layer menu's Add Layer options - both reach the same dialog.
Point it at your CSV or text file and QGIS shows a live preview of the parsed table at the bottom of the dialog. If the columns in that preview look wrong - everything crammed into one column, headers treated as data - adjust the format options: QGIS handles standard CSV, custom delimiters (semicolons, tabs, pipes), and even regular-expression delimiters for awkward files. The preview is your friend; do not proceed until it shows clean columns.
Under the geometry settings, choose Point coordinates and set the X field to your longitude or easting column and the Y field to your latitude or northing column. QGIS also offers a DMS coordinates option for fields that contain degrees-minutes-seconds notation - useful, though it expects consistently formatted DMS, not a mixture of styles.
Finally - and this is the step that matters most - set the Geometry CRS to the system your data is actually in: EPSG:4326 for GPS-style lat/long, EPSG:27700 for British National Grid, the correct zone code for UTM or State Plane. Click Add and your points appear, with QGIS reprojecting them on the fly to whatever CRS your project canvas uses.
One behaviour worth knowing: a delimited text layer stays linked to the source file and is not directly editable. To turn it into a permanent, editable dataset, right-click the layer and export it to a real format such as GeoPackage or shapefile.
Importing coordinates into ArcGIS Pro
ArcGIS Pro's equivalent workflow is the XY Table to Point geoprocessing tool (in the Data Management toolbox). You can run it directly from the Geoprocessing pane, or add your CSV to the project first and right-click the standalone table in the Contents pane to reach the same operation via Display XY Data.
The tool asks for the input table, the X Field (longitude or easting), the Y Field (latitude or northing), an optional Z field for elevation, and the Coordinate System of the input coordinates. As in QGIS, that coordinate system parameter describes what the numbers already are - GCS WGS 1984 for GPS lat/long, the specific NAD83 State Plane zone for US survey data - not the system you want the output displayed in. The output is a proper point feature class in your project geodatabase, immediately usable for selection, editing, and analysis.
If you are maintaining legacy ArcMap projects: the old route was Add XY Data, which created a temporary "XY event layer" that had to be exported to a feature class before serious use. ArcGIS Pro's XY Table to Point skips that intermediate step and writes a feature class straight away. ArcMap is retired and no longer supported by Esri, so new workflows should be built in Pro.
The concept that trips everyone up: data CRS vs map CRS
Both QGIS and ArcGIS maintain two separate ideas of "coordinate system", and confusing them is the root of most import problems.
The layer (data) CRS describes the numbers stored in your file. The project or map CRS describes how the canvas draws everything. They do not need to match: both packages reproject layers on the fly, so a WGS84 point layer displays perfectly on a British National Grid map - provided the layer's CRS is declared correctly. On-the-fly reprojection is a translation service, and like any translator it can only work if it knows what language the source is speaking.
This is why "assigning" a CRS and "reprojecting" to a CRS are different operations. Assigning changes the label - the metadata that says what the numbers mean - without touching the numbers. Reprojecting mathematically transforms the numbers from one system to another. If your points are in the wrong place, assigning a different CRS until they look right is a gamble; the correct fix is to find out what system the data was genuinely captured in and declare that. When the source and target sit on different datums - WGS84 GPS data going into an OSGB36 project, for example - the software also applies a datum transformation, which is why declaring the right source CRS matters down to the datum, not just the projection.
The classic failure modes (and what each one means)
Imported points that misbehave almost always fail in one of a few recognisable ways. The symptom tells you the cause.
- Points in the ocean off West Africa. The infamous "null island" at 0°, 0° in the Gulf of Guinea. Your X or Y values parsed as zero or empty - usually blank cells, text that failed to parse as numbers (stray units, thousands separators, degree symbols), or a decimal-comma file read with a decimal-point parser. Fix the source data, not the map.
- Points mirrored into the wrong hemisphere or wrong continent. X and Y are swapped - latitude went into the X field. A UK dataset that plots near Somalia (lat and long transposed around the 0° meridian region) or US data landing in Antarctica are the classic signatures. Re-run the import with the fields the right way round.
- All points in a tiny cluster near the origin of a projected map. Degree values (lat/long) were imported with a projected CRS declared, so numbers like 51.5 were read as metres. Declare EPSG:4326 (or whatever geographic CRS the data is truly in) instead.
- Points offset by a consistent 50-150 metres. A datum mismatch - typically WGS84 data declared as OSGB36 or vice versa, so no datum transformation was applied. See Why Are My Coordinates Offset or Shifted?.
- No points at all, or fewer than expected. Parsing failed silently on some rows: a semicolon-delimited file read as comma-delimited, European decimal commas, a UTF-8 byte-order mark corrupting the first header, or header rows mixed into the data. Check the import dialog's preview (QGIS) or the table (ArcGIS) row by row.
Null island is a data problem, not a CRS problem
The format problem: GIS import tools expect clean numbers
Both import dialogs share an assumption that real-world data routinely violates: that every coordinate is already a clean pair of plain numbers. Decimal degrees like 51.5074, or projected easting/northing like 530029. One format, one column each, no decoration.
Field data rarely arrives that way. A typical inherited spreadsheet mixes DMS and decimal notation between rows, wraps values in quotes, appends units or hemisphere letters, uses European decimal commas in half the file, or stores the whole coordinate in a single column with inconsistent separators. QGIS's DMS option helps with uniformly formatted DMS, but neither package will parse a genuinely mixed or messy file - rows fail silently or the import aborts.
The reliable approach is to standardise the coordinates before the GIS sees them: convert everything to clean decimal degrees (or a clean projected easting/northing), one value per column, then run the import. That preparation step is exactly what CoordinateMapper is built for.
Preparing your data with CoordinateMapper
To be clear about the division of labour: QGIS and ArcGIS are full GIS packages with their own reprojection engines - once your data is in, they can transform it between any CRS you like. CoordinateMapper is not a replacement for either. It earns its place one step earlier, where the GIS import tools are weakest.
Paste or upload your coordinates in whatever state they arrived - mixed DMS and decimal, UTM, MGRS, UK Grid references, State Plane with FIPS zones, compact or oddly punctuated notations. The parser auto-detects each format, applies the correct datum handling, and plots every point on a map so you can visually confirm the data is sane before it goes anywhere near your project - a thirty-second check that catches swapped columns and rogue rows far earlier than a failed import does. Then export a clean CSV of decimal lat/long (or KML), and the QGIS or ArcGIS import becomes the one-minute job it should be.
It is also simply faster for the small jobs: a handful of DMS coordinates from an email do not justify opening a GIS, and State Plane batches that need converting to lat/long before wider use are a paste-and-download exercise. Single conversions and CSV/KML export are free; batch file processing and DXF export for CAD are part of Pro.
Frequently asked questions
- Q: Why did my imported points end up in the ocean near Africa?
A: Points at 0°, 0° ("null island") mean the X or Y values parsed as zero - blank cells, unparseable text, or a delimiter/decimal-separator mismatch in the CSV. Points elsewhere in the wrong ocean usually mean swapped X/Y columns or a wrong CRS declared at import. Check the source rows first, then the field assignments, then the CRS. - Q: What CRS should I choose when importing?
A: The CRS the data is actually in - not your project's CRS. GPS and phone exports are almost always WGS84 (EPSG:4326). UK easting/northing is British National Grid (EPSG:27700). UTM and State Plane need the correct zone-specific code. If it is genuinely unknown, the magnitude of the numbers narrows it down: small decimals are degrees, six-digit metre values are projected. - Q: Does column order matter - X/Y vs lat/long?
A: The column order in the file does not matter; the field assignment does. X = longitude or easting, Y = latitude or northing. Because the spoken convention is "lat, long" (Y first) while GIS asks X first, this is the most common swap. If your points mirror to an impossible location, this is the first thing to check. - Q: Can QGIS import DMS coordinates directly?
A: Yes, within limits. The Delimited Text dialog has a DMS coordinates option that parses consistently formatted degrees-minutes-seconds fields. It will not handle a file that mixes DMS with decimal degrees or uses erratic punctuation - convert those to clean decimal degrees first. - Q: How do I import a CSV with easting and northing instead of lat/long?
A: Exactly the same workflow - easting goes in the X field, northing in the Y field - but the CRS must be the projected system the values belong to (EPSG:27700 for British National Grid, the zone-specific code for UTM or State Plane). Declaring EPSG:4326 for metre values puts everything wildly off the map. See the Easting/Northing converter if you would rather convert to lat/long first. - Q: My points display fine but are offset from my basemap by about 100 metres. Why?
A: That is the signature of a datum mismatch - most commonly WGS84 data declared as OSGB36 (or vice versa), so no datum transformation was applied during reprojection. Confirm which datum the data was captured in and re-declare the layer CRS correctly.
Takeaway
Importing coordinates into QGIS or ArcGIS is a two-decision job: assign X and Y to the right columns, and declare the CRS the data is genuinely in. QGIS does it through the Delimited Text layer, ArcGIS Pro through XY Table to Point, and both reproject flawlessly once those two declarations are correct. Nearly every failed import traces back to messy source data or a wrong declaration, not the software - so clean and standardise the coordinates first, verify them visually, and the import itself becomes the easiest step of the day. The same preparation pays off again if the data is also headed for CAD: see How to Import Coordinates into AutoCAD.
Arriving with messy coordinates?
CoordinateMapper cleans mixed-format coordinate data - DMS, decimal degrees, UTM, State Plane, UK Grid, stray labels and all - and exports it as a clean CSV or KML that QGIS and ArcGIS import without a single parsing error. Batch file processing and DXF export are part of Pro; single conversions and CSV/KML export are free.
See what Pro includes