Plot CSV, Excel & TXT Files Online

Drop the file onto the plotter at the top of this page or click “Choose a file”. The first chart appears as soon as the file is read, with a time or number column on the x axis and the first numeric column as a line. Pick other columns from the toolbar or by clicking the column headers. There is nothing to install and the file is not uploaded.

Open the .xlsx or .xlsm file here and your browser reads it on your own device. The plotter opens the first sheet with data and finds the header row, even below a report title; switch sheets from the menu in the top bar. Dates, times, percentages and the saved results of formulas are read as values. Old .xls files need to be saved as .xlsx in Excel first.

Yes. Files far beyond Excel’s limit of 1,048,576 rows open in seconds: in our tests in Chrome, a 2-million-row CSV took about 1 second, and a 10-million-row CSV opened and drew its first chart in about 3 seconds (measured in Chrome on a recent laptop, with a 423 MB CSV of 10,000,000 rows and 5 columns). Peaks and dips stay visible in the overview, and when you zoom in the full detail reloads instantly.

Yes. Move any series to the right axis with the L / R button in the toolbar or from the column header menu, and it gets its own scale. That is how you show temperature and pressure, or units and revenue, on one chart. Up to eight series fit on the two axes.

Yes. Use stacked lanes to give each signal its own Y axis and range, with one shared time axis below. The signals sit in separate lanes, so this is not an overlaid dual-axis chart. Zoom in time to explore the lanes together.

Because no column is the X yet, so each row is plotted at its position in the file, evenly spaced. The line at the top of the column list then says “X: Row number”; press “change” next to it and pick your time, date or number column. When a column would make a better X, such as a sorted frequency column, the plotter can suggest it under that line, and Use sets it.

Make the frequency column the X: press “change” in the column list’s X line and pick it, or choose it as the X axis column in Plot settings → Axes. Then tick “Log scale, X” in the same Axes tab, and each decade, such as 1 to 10 Hz or 100 to 1,000 Hz, gets the same width.

A log axis has no place for 0 or negative numbers, since the logarithm of 0 is minus infinity. With “Log scale, X” ticked, rows whose X is at or below 0, such as the 0 Hz row of a spectrum, are left out of the chart and a note in its corner counts them. They stay in the file and the data table, and unticking the box draws every point again.

No. The file is read and charted by your browser on your own device, and neither the file nor its name is sent anywhere. You can check this in the Network tab of your browser’s developer tools while a file loads.

Yes. The Math button turns formulas (for example the magnitude of three axes), low-pass, high-pass and band-pass filters, moving averages and integrals into new channels you can plot like any column. It also opens FFT and PSD plots with a Grms readout, transfer functions with coherence, shock response spectra, rainflow cycle tables and statistics. Everything is calculated on your device, your math channels come back when you open the file again, and our results are checked against SciPy.

Line, step, bar, area and scatter charts, and histograms, all zoomable with the scroll wheel, a pinch or a drag (Shift+scroll for the Y axis), plus a crosshair with live values and PNG export. Any chart can also summarise its rows: sales by month, quarter or year, totals by product or region (side by side or stacked), the average of a signal every 10 seconds, or the time spent in each speed band. A file of sales or orders opens as a monthly bar chart, and the summary can be downloaded as a CSV. Box plots show the median, quartiles, whiskers and outliers of each series side by side, over every row, with their numbers as a CSV.

Yes. Choose or drop several files at once, or use “Add file”, and plot channels from each of them together: run A against run B, before and after a change, or this year’s sales against last year’s. Runs that started at different times can be lined up at their start, and the files can be compared side by side in tables with the differences worked out for you. Up to eight files can be open together, and your plots come back when you open the same files again.

Yes. Many data loggers and historians export every signal in one column, with a time, the signal’s name and its value on each row. The plotter recognises these files and splits them into one channel per signal as they open, each with its own timestamps, so signals recorded at different rates plot together correctly, even with thousands of signals. Search and tick the signals you want, zoom, compare them, read their statistics or calculate across them. “Keep as rows” shows the file exactly as it is, and “Download as wide CSV” puts the signals side by side, one column each, ready for Excel.

Open the Reset page at megarows.com/reset and choose “Reset now”. The layouts, math channels and settings this browser remembers go back to the defaults; your light or dark theme is kept. Your files are never stored, so nothing in them is lost. In the plotter, “Reset everything to defaults…” at the end of “Do more” does the same. And if a saved layout ever stops the page, the plotter opens that file with the default layout next time and offers to try the saved one again.

CSV, TSV, TXT and DAT files with any common separator (comma, semicolon, tab, pipe or space-aligned columns), Excel workbooks (.xlsx and .xlsm), JSON or JSON Lines, MATLAB .mat files from v4 to v7.3, including Simulink- and dSPACE-style logs, Parquet files, HDF5 and NetCDF-4 files, and NI TDMS files from LabVIEW, DIAdem and DAQmx loggers; any of them but TDMS also compressed with gzip (.csv.gz) or inside a ZIP archive. The format is recognised from the file’s content, not only its name. MF4 measurement files are planned for later.

Yes. Choose an .h5, .hdf5 or NetCDF-4 .nc file and the first table is plotted; “Variables” shows every group and dataset with its type, shape, first values and attributes, and a tick plots a dataset. Datasets of a group that have the same length open as one table against their time axis (a dimension scale, a dataset named like a time, or dates and times from units such as “hours since 2026-01-01”); 2-D datasets become columns, named from a dimension or an attribute when the file has them, and record arrays a table, with units from the units attribute. Packed values are unpacked and fill values left empty. Datasets of three or more dimensions and NetCDF 3 (classic) files can’t be plotted yet. The file is read on your device and never uploaded.

Yes. A file compressed with gzip (.csv.gz, .tsv.gz, .json.gz and the like) opens like the file inside: it is unpacked on your device as it is read, and the file line shows both sizes. A ZIP archive with one data file opens it straight away; with several, a short list shows each file with its type and size: tick the ones you want, up to 8 at once, and each opens as a file of its own, ready to compare on one plot. “Add file” opens more next to them. Password-protected archives can’t be opened; extract them first.

Files from Excel 97–2003 use an old binary format that MegaRows does not read. Open the file in Excel or LibreOffice, choose Save As → Excel Workbook (.xlsx), and open the new file here. Password-protected workbooks cannot be opened either, so remove the password in Excel first.