Hi @User195 — Max and I are discussing methods in other topics and we have somewhat different (but perhaps not incompatible!) ways of thinking about challenges like yours. Max is advising on statistical issues, which are especially relevant since you opened the topic with a request for a tool that would offer web-based support for statistical operations. My advice would be to go back a few steps and work on the question formulation first, before moving to data gathering and statistical analysis. As an exercise that may help, I would start with the capacities of Zenobase in your mind and think about a project design whose analytical phase used only Zenobase affordances and nothing more. The reason I suggest this (perhaps absurd-seeming) constraint is that @ejain, who wrote and maintains Zenobase, is a very experienced self-researcher and is a member of the QS Forum, so you will not only be starting with a tool that has been made based on a lot of practical knowledge, but its creator may be able to discuss issues with you here.

So, what does this constraint mean? In general, it means you’ll be working with pretty simple time series plots. So your questions should probably not be too complex. You won’t be looking at large numbers of variables and trying to pick up relative effects over arbitrary time windows accounting also for delayed effects and rhythmicity, for instance. Instead, you’ll be making some reasonable guesses about what might affect your skin and then collecting observations for some time that will allow you to generate new ideas and also refine the ideas.

The most important thing to track, if you aren’t already, is a measure of the skin irritation. Do you have a quantitative measure of this going already? Have you tested it in practice; that is, are you able to confidently evaluate the severity of the condition using your scale of measurement without causing too much friction in your daily life? If you’ve been doing this for a while, you’ve already made a great step forward! The next step, which you may have already done, is just to plot this single variable to get a feel for the typical range, the frequency and severity of the more extreme measures, and cyclical factors. Just looking at this plot (which Zenobase can make, but so can any spreadsheet program almost), can generate some good ideas to test.

Third step, would be to think about the second or third data source you’d like to use. I’ll stop here in case my advice is misguided. What do you think?