Topsoil Design Ideas

[noahmclean]

Hi Topsoilers,

I've started a google doc with some top-level design ideas for how isotope geochemists and geochronologists might interact with Topsoil. Find it here: https://docs.google.com/document/d/1thXblPCvkIMTIyKZRAEtlbTFZGTFA1oCWRD1bNVqzhc/edit?usp=sharing

An intuitive and interactive interface with isotope system-specific inputs, controls, and outputs means that beginner through expert users can ask better questions, and better questions means better answers.

These are just a few of my ideas, typed out quickly a couple weeks ago. I think the next stage in the top-level design needs to one or more groups of users coming together to brainstorm about what they want out of a next-generation data visualization and calculation engine.

[spenchristoph]

Hi Noah, This is an impressive wishlist. Might I add the x,y,c plot where bivariate data are colored on a spectrum using a third variable?

On Sat, Jun 28, 2014 at 11:21 AM, Noah McLean notifications@github.com wrote:

Hi Topsoilers,

I've started a google doc with some top-level design ideas for how isotope geochemists and geochronologists might interact with Topsoil. Find it here:

https://docs.google.com/document/d/1thXblPCvkIMTIyKZRAEtlbTFZGTFA1oCWRD1bNVqzhc/edit?usp=sharing

An intuitive and interactive interface with isotope system-specific inputs, controls, and outputs means that beginner through expert users can ask better questions, and better questions means better answers.

These are just a few of my ideas, typed out quickly a couple weeks ago. I think the next stage in the top-level design needs to one or more groups of users coming together to brainstorm about what they want out of a next-generation data visualization and calculation engine.

— Reply to this email directly or view it on GitHub https://github.com/CIRDLES/topsoil/issues/57.

[johnzeringue]

@noahmclean, to add to that last bit, one of the most important pieces of the brainstorming phase is prioritization of features. That is, knowing as little about geology as I do, it's clear to me that Noah's list of ideas is a mix of stretch goals, necessities, and things in between. However, I'm sure that you all are much better at deciding what's important and what's not that we are. While adding new ideas is always appreciated, it's equally important, if not more so, to make clear the features that you need before being able to use Topsoil and the features that would greatly enhance your work or user experience. Detailed use cases are great for conveying this to the dev team.

If I might make another request, we're doing a lot of behind the scenes work for chart creation right now, and it would be great if you could give us images/mockups/descriptions of anything that we've yet to see. For example, @spenchristoph, I vaguely understand what you're getting at, but specific examples (especially images) allow us to be well prepared to implement this stuff.

[spenchristoph]

@johnzeringue, attached is an image that displays several things I would love to see in Topsoil. 1) KDE's plotted along the x-y axis 2) a third variable plotted along a color spectrum 3) KDE plotted along the spectrum for the third variable

[johnzeringue]

Is the KDE scaled to anything in particular (with regards to its "height") or is the shape all that matters?

[bowring]

Check out the KDE code in Redux for some ideas.

On Sun, Jun 29, 2014 at 4:20 PM, John Zeringue notifications@github.com wrote:

Is the KDE scaled to anything in particular (with regards to its "height") or is the shape all that matters?

— Reply to this email directly or view it on GitHub https://github.com/CIRDLES/topsoil/issues/57#issuecomment-47473197.

[jwboyce]

My use of Topsoil would fall into two categories: "geochronology" and "not geochronology". So I'm pretty much just commenting to say "thanks" for considering all the non-geochron applications in the design of Topsoil. As a recovering geochronologist, I think many geochemists would benefit from the kind of rigorous treatment of uncertainties that the geochronology community takes for granted. I also spend quit a bit of time in ternary space, and so I'm pleased to see that is on the list as well.

[ryanickert]

Hi Noah (et al.),

I think the overall structure of Noah's document (BTW, is it supposed to be community editable?) is great.

I would add one visualization feature that I don't think many are aware that they might want - two dimensional KDE/PDE plots (e.g., doi: 10.1029/2005GC001052). I have been doing this in Matlab to simplify large, heteroscedastic common-Pb datasets. It is quite a bit trickier because the KDE/PDE needs to be contoured in order to be displayed on a 2D plot, but I think that this is a powerful visualization tool for large datasets (especially visualizing outputs from databases or compilations).

I would also emphasize that we need to be able to save a particular configuration for a plot (or set of plots). When I produce a plot for a manuscript or a talk, I often will find myself either returning to the plot to make small alterations, or I have a different dataset that I want plotted in exactly the same way. They may be different tiles on a multi-plot figure, for example. I would like to be able to save and recover everything that made that plot unique (size, line-width, colours, scale etc.) so that I can exactly reproduce it.

[jwboyce]

Ditto EVERYTHING Ryan Ickert just wrote.

[noahmclean]

Great comments and suggestions all, thanks for the input! @spenchristoph, @ryanickert, and @jwboyce, I've added your suggestions to the document and opened it up for editing by anybody with the link above. All, feel free to add or comment in the doc, and please make a note of it here at github so we can keep the discussion in one place. Keep the suggestions coming!

[mcbuzzaldrin]

Although we are not very knowledgeable in geology in general we found the UI and even the wiki quite confusing especially for someone with little computer experience. We also had trouble differentiating between why people should use Topsoil over Isoplot when it seems to be an Isoplot derivative?

After downloading the software we brought to a blank page. While very simple, it's a little too simple as it leaves the user to figure out the system although it does provide an option to paste from the clipboard. Most people who are not computer scientists will not search to read a wiki before using the software. We think it will be more useful for the tutorial to be within the software or at the very least help page. One user also wished the were more options to interact with the visualization created.

We find that once people are used to Isoplot they will not have a great reason to switch as as far as we see the design does not provide the user easier usabillity or have options that Isoplot doesn't have.

[johnzeringue]

@mcbuzzaldrin, although there are clearly shortcomings in the UI, it was designed to be similar to Excel, which all users of Topsoil are assumed to be familiar with.

There's already an open issue (#127) to address initial usability, along with ideas from @jthevos and @winslowdibona.

Isoplot is an excellent piece of software that will likely retain users for many years. For now, Topsoil currently aims to mimic Isoplot in a number of ways, while still differentiating in the following ways:

• open source
• open development
• active development (Isoplot is no longer maintained)
• being future-proof
• being platform-independent
• using vector graphics

Thanks for commenting, though. There's certainly some truth to it all at the moment, but we love getting feedback!