Not that I'm saying Excel's graphing is excellent, but the graphs at https://datanitro.com/blog/assets/img/chart3.PNG are a bit of a cheat since there's very little customization done to the Excel graph.
Another thing worth mentioning is how the plots were made. Here, the author seems to have just taken screenshots from MS Excel, and not actually exported anything to a portable format. Even the jagged, choppy Excel graphs look fairly smooth when exported as a .PDF file (as Excel does save the vector data for smooth curves) or another comparable vector format.
Specifically, in the author's first example (the exponentially decaying sinusoid), viewing that graph in MS Excel will always appear jagged and pixellated. However, if you were to export the chart as a .PDF, the aliasing artifacts would completely disappear (or at least be comparable with the Python-rendered example). I use this technique when I make some quick charts/graphs in Excel and want to include them in a LaTeX document. You can move a chart/graph to it's own full-page "worksheet", and then Save As a PDF file. Since the PDF output is in vector format, it works nicely with LaTeX documents (as you can use \includegraphics{} natively with .PDF documents), and scales well to any size without artifacts.
That being said, there is one thing I do agree with here: the way MS Excel renders graphs/charts when editing them is pretty horrible. You can, however, get around the issue by choosing an appropriate output format (you might just have to zoom in a lot when editing to avoid the aliasing temporarily).
> That being said, there is one thing I do agree with here: the way MS Excel renders graphs/charts when editing them is pretty horrible. You can, however, get around the issue by choosing an appropriate output format (you might just have to zoom in a lot when editing to avoid the aliasing temporarily).
The thing is, there's no excuse for that. Excel graphs should be good by default. Same with the OP's issue. The amount of time I've wasted pretty-ing up an Excel graph is crazy. DataNitro looks great. Shame it's Windows only though...
The DataNitro graph also omits all the empty space on the vertical axis below 300, which makes for a more informative graph. The Excel graph does not, even though there's no technical reason Excel would prevent you from doing so.
Which means that the two graphs being compared aren't really the same, making this more of a comparison of "chart made by someone who knows how to make a good chart vs. chart made by someone who doesn't" than a comparison of DataNitro vs. Excel.
The Python graph is more useful if the story you want to tell is that the values vary a lot. Depending on the data being plotted, that could be either more informative, or more lying. For example, suppose that y values varied between 2999.996 and 3000.004. Would you want to see a y axis range from 2999.995 to 3000.005?
I think it is OK for Excel to prefer "make sure you aren't lying" over "show as much of the variation as you can". It is not as if it is hard to change the plot range in Excel.
The real advantage to plotting via code isn't just better visual styles, but mass plot generation. The process in Excel:
1. select data
2. select chart type
3. customize chart
4. export chart (for Word doc inclusion)
can be very time-consuming repeated with dozens of charts. Each time someone wants to alter the style (a manager's arbitrary decision to use three decimal points instead of two etc.), steps 3 and 4 must be repeated for each chart. Or data get corrected in a way that alters all values and everything needs to be re-exported, and the axes need to be fixed. These "little" items really eat into time as data are massaged and report drafts get reviewed and corrected.
That may be intentional, but needs to be noted.