The approach I usually use for CSV data in JSFiddle examples is
a. Put the data in a <pre> block at the end of the HTML mark-up, usually with the id “data”.
b. Add pre {display:none;} to the CSS.
c. Replace the d3.csv(filename, callback) function call with a d3.csv.parse(text) call, using the text content of the <pre> block as the input to the parse function.
Because the parse function doesn’t use a callback, it just returns the data array, you need to save that output in a variable of the same name as your callback data parameter.
In other words, if your example looks like
d3.csv("data.csv", function(error, data) {
if(error){console.log("Could not read " + "data.csv");
/* Do something with `data` here */
});
The JSFiddle-friendly version would look like:
//d3.csv("data.csv", function(error, data) {
// if(error){console.log("Could not read " + "data.csv");
var data = d3.csv.parse( d3.select("pre#data").text() );
/* Do something with `data` here */
//});
If you would rather have a full working example that uses the file-reading methods as intended, there are other options as mentioned in the comments. Tributary also allows external data files I think.