Search code examples
pythonhtmloutputtabular

Pros and Cons of html-output for statistical data

I am using Python3 to calculate some statistics from language corpora. Until now I was exporting the results in a csv-file or directly on the shell. A few days ago I started learning how to output the data to html-tables. I must say I really like it, it deals perfect height/width of cell and unicodes and you can apply color to different values. although I think there are some problem when dealing with large data or tables.

Anyway, my question is, I'mot not sure if I should continue in this direction and output the results to html. Can someone with experience in this field help me with some pros and cons of using html as output?

Solution

  • The question lists some benefits of HTML format. These alone are sufficient for using it as one of output formats. Used that way, it does not really matter much what you cannot easily do with the HTML format, as you can use other formats as needed.

    Benefits include reasonable default rendering, which can be fine-tuned in many ways using CSS, possibly with alternate style sheets (now supported even by IE). You can also include links.

    What you cannot do in HTML without scripting is computation, sorting, reordering, that kind of stuff. But they can be added with JavaScript – not trivial, but doable.

    There’s a technical difficulty with large tables: by default, a browser will start showing any content in the table only after having got, parsed, and processed the entire table. This may cause a delay of several seconds. A way to deal with this is to use fixed layout (table-layout: fixed) with specific widths set on table columns (they need not be fixed in physical units; the great em unit works OK, and on modern browsers you can use ch too).

    Another difficulty is bad line breaks. It’s easy fixable with CSS (or HTML), but authors often miss the issue, causing e.g. cell contents like “10 m” to be split into two lines.

    Other common problems with formatting statistical data in HTML include:

    • Not aligning numeric fields to the right.
    • Using serif fonts.
    • Using fonts where not all digits have equal width.
    • Using the unnoticeable hyphen “-” insted of the proper Unicode minus “−” (U+2212, −).
    • Not indicating missing values in some reasonable way, leaving some cells empty. (Browsers may treat empty cells in odd ways.)
    • Insufficient horizontal padding, making cell contents (almost) hit cell border or cell background edge.

    There are good and fairly easy solutions to such problems, so this is just something to be noted when using HTML as output format, not an argument against it.