Writing a no-code drag-and-drop tools for transforming data (join, filter, reformat, reorder etc). The hardest bit is handling cascading changes to the column structure (e.g. removing and reordering columns) intuitively - particularly if they change the input file to one with less columns or differently ordered columns.
Love this idea. At one point a few years ago I was keen on creating a visual interface for lodash for non-coders. Watching nontechnical users try to navigate repetitive tasks is always hugely frustrating. Decided in the end that if they won’t use a spreadsheet they probably wouldn’t use a website.
There's already IBM's DataStage. I currently use this for ETL work at my job, although we're working on moving our ETL work to Python and an in-house framework due to licensing costs.
I know there are various Extract Transform Load tools aimed at professional data scientists. My product is aimed at numerate professionals who aren't programmers or data scientists and have never heard of 'ETL'. The idea is that they can install Easy Data Transform, transform their data and output it in a few clicks, without programming. It is also very cheap compared to many commercial ETL tools.