CSV Quick Chart Generator
Drop a CSV file and see a chart before you've had time to open Excel. Pick your columns, choose a chart type, and download the result as a PNG. Nothing ever leaves your browser.
Drop your CSV here
or click to browse · CSV, TSV, or TXT
Column Statistics
Learn more: reading the stats behind your chart
How outliers get flagged
Outlier marking runs on scatter charts only, and only while "Color by" is left at (none). The tool takes the mean and standard deviation of the Y column, then flags every point sitting more than two standard deviations from that mean. Flagged points move into their own red "Outliers (>2σ)" series so you can see them without losing them from the plot.
It is a blunt rule, and small columns expose that. Take the values 10, 12, 11, 13 and 40: the mean is 17.2 and the standard deviation 11.44, putting the threshold at 22.89, while the 40 sits 22.8 away from the mean. The extreme value inflates the spread it is being measured against, so it escapes the flag by a hair.
Below four numeric values the check is skipped entirely, since a mean and spread from three points say almost nothing about what counts as unusual.
What the trendline is actually fitting
The trendline is ordinary least squares: the straight line whose squared vertical distances to your points are as small as possible. The R² printed beside it in the legend is the share of the column's variation that line accounts for, 1 being a perfect fit and 0 meaning the line explains nothing at all. Fits worse than a flat average are reported as 0 rather than a negative number.
What it fits depends on the chart type. On a scatter chart it regresses Y against the real X values. On a line chart it regresses against row position instead, treating the points as evenly spaced 0, 1, 2 and so on, because the X axis there is a list of labels rather than a numeric scale.
That distinction matters if your dates have gaps. Twelve readings taken monthly and twelve taken at scattered intervals produce the same fit positions, so the slope you get is per row, not per day.
Where the column stats come from
The stat cards read the raw column rather than whatever ends up on the chart. Aggregation, bar sorting and the row filters all leave them untouched, so a filtered chart still sits under stats describing the whole file. Blank cells are counted separately as nulls and left out of the maths instead of being treated as zero.
A column is only treated as numeric if every non-blank value in it parses as a number. A single "n/a" or a thousands separator reclassifies the whole column as text, which is the usual reason a column you meant to chart never appears in the Y picker.
FAQ
Should I delete flagged outliers before charting?
Not automatically. A flagged point can be a data entry mistake worth fixing, or it can be the most interesting real value in the set. Check it against the source before deciding either way.
Why does the median differ from the mean in my column stats?
The mean is pulled toward extreme values and the median is not. A wide gap between the two usually means the column is skewed by a handful of unusually high or low values, worth a look before you trust the average as a summary.
Why is there no trendline option on my bar chart?
The trendline toggle only appears for line and scatter charts. Fitting a slope across bars whose order you can change at will would not mean anything, so bar, pie, donut and histogram views leave the option out.