What inspired this wish list request: In my field, electronic notebooks (ELN) seem to be everywhere. While they bring critical benefits, the few that I've used don't seem well set up for experimentally designed projects. It's very much "what did you do" first, and data second. With a DoE, to use the lab notebook correctly, depending on what software you have, you might have to do a whole new entry for each experiment, an entirely unwieldy premise that deters one from using ELNs correctly.
On the flip side, especially with projects, JMP invites you to include all of your data because it is so straight-forward. However, one of the lackings is that while I want to store as much info as I can in JMP, I can't store my non-numerical data, not easily at least. Currently I'm sitting on hundreds of photos that I took of different steps in my experiment. I could put them in the data table, but that would become cumbersome extremely fast. Aside from the annoyance of having to deal with even more columns than the 200+ I already have, during some experiments I took one picture for a given step, while during others I took 5. Given that there's no organizational principle to how I took the pictures, some picture columns would be sparse, while others dense: an underutilization of JMP's great column functionality.
Additionally, and maybe more importantly, there is the general problem of proper transfer of knowledge, especially in academia where there is high turnover of students. Many of the lessons learned come not from the data, but seeing what was noted at the time of experimentation. However, with lab notebooks, just going through the lab notebook from some random person you never met from years ago on a project you don't know can be both daunting, and sometimes counterproductive. In my experience, it almost never happens. ELNs can make this process much easier because of their search features, uniformity of input, etc. However, the problems with ELNs stated before remain, especially in that they are often optimized for "what did you do" rather than the data and analysis. JMP has virtually no way to do this, except to upload a picture of a lab notebook into a project file, which would be as likely looked at as the physical pages in the lab notebook.
What is the improvement you would like to see: I think a few things would be very useful for researchers:
- Associate photos with rows and columns: Allow a set of pictures to be associated with a particular column or set of columns, each picture being assigned to a row (or set of rows if the same setup use used for multiple experiments, such as in a split-plot experiment). Notation and maybe even image markup (e.g. drawing an arrow to an important part of an instrument, or, in the case of an image associated with multiple columns, assigning parts of the image to different columns) would also make sense here.
- Quick access of the experimental photos: This is rather self-explanatory, but I'm not sure the best way. Perhaps have a view mode that draws small outlines of the regions that are associated with a particular image? This wouldn't work though with tons of overlapping regions if there were any. And then possibly a persistent hover window of the photos that you are interested in that you x out of, can drag around, resize, and scroll through that you could access from clicking on a cell, column, or row, with the option to do a pop-up or new file if using JMP projects.
- Save images of notebook pages with metadata: Have images of notebook pages stored in a way that relevant sections can be accessed while browsing data. Something like NVivo would be overkill, but a simple click and drag a box, polygon, or freeform line around the important area of a lab notebook and associate with individual/set-of cells/rows/columns. I'm envisioning something like this: I'm looking at the JMP table for some experiment, see some weird outlier in row x, click some button or press a hotkey, the lab notebook page pops up, and I can read what was written when the experiment was performed, and ah! that was the day before we found the clog in instrument Y, which was used in this experiment. Now I have reason to suspect this outlier is due to instrument failure rather than a true outlier. Could this be done by re-reading the physical lab-notebooks? Sure! Would it be a thousand times easier to not have to go back and forth between the physical copy, which may have pertinent info across multiple pages, and the computer screen, but rather have associated sections available next to the data on screen? YES!
- Related Idea
- One of the issues with JMP is when you have a gagillion columns. Lots of rows is manageable, but tons of columns can get confusing fast. And while column grouping is very helpful, a more general overview would fit well into what I've said above. A common tool, especially in business, is a process map. In STEM, a process map could summarize a variety of things: a manufacturing process, a mechanism, an experimental procedure. If process mapping were implemented, then a very complex JMP table could be given an overview, where different columns are associated with different steps in the map. For example, if the map represents an experimental procedure for a chemical reaction, the concentrations, temperature/pressure, reaction vessel size, etc could be associated with one step, while the crystallization condition columns could be associated with step two, and the final yield could be associated with the third (analysis) step. In my mind, I'm envisioning zooming out from the data table, and a process map fading in, with associated column names listed under each step in the map and maybe even pre-scripted analyses, but I'm sure there's a better way to do it.
- Because process mapping is a "related idea" and not the main one, I'll put the "why this idea is important" here: While JMP journals and dashboards have some aspects that touch on these ideas, having an overview of a process, experimental procedure, in a graphical, interactive format that's connected to the data, either in the data table or in a journal, would be very helpful for both the experimentalist's brain-space, especially with many-column'd tables, as well as in communication of results, especially with people in a meeting or conference unfamilliar with a process.
Why is this idea important: Because of its easy-to-use and feature-rich, JMP (rightly!) invites the user to make JMP the hub for data exploration and analysis. However, the process of experimentation and collecting data is just as important as the data itself, because a poor experiment will generate poor data. Poor communication of what you did will also lead others to have poor experiments.... leading to poor data. Adding functionality to store your non-numerical experimental info in an easily referencable way would very much help organization, analysis, and tech transfer.
Note:
There are a couple of related wish-list items that readers should go upvote: