Random Charts & Nonsense

the addict · Tennis legend

Random Charts & Nonsense

AliBlahBlah · All-time great

Still trying to figure out CH - adding just a note seemed to dissociate it form it's position in the chart, I couldn't 'see' where CH fell just from reading it and trying to imagine it.

But, in the meantime, from related data, here's this:

The two unlabelled marks are Katie SWAN (at left CH ~180) and Emily WEBLEY-SMITH (at right CH ~240)

And, a very rough look at the version that has every ranked Brit woman ever - the currently ranked are in gold here

foobarbaz · Strong Club Player

I just had a look at the chart showing GB Women's Rank Group with Time. Instead of using a single icon type at each data point, might it be useful to show either a directional arrow (up/down) or a solid dot, to indicate how a player has progressed (in rank) from the previous rankings set? Apologies, if this has already been suggested/discounted.

indiana · Tennis legend

AliBlahBlah wrote:
Still trying to figure out CH - adding just a note seemed to dissociate it form it's position in the chart, I couldn't 'see' where CH fell just from reading it and trying to imagine it.

[ ... ]

I tend to disagree and think it a pretty good mix of the graphical and tabular. I can see folk looking at the graph, which I do really like as is, and the note of the CH just to the side. It doesn't have to be actually contained within the graph to be associated.

If you find your elegant solution good but personally my eyes don't take to too much in a graph, call it clutter or whatever. I like to see and get it fairly quickly.

AliBlahBlah · All-time great

foobarbaz wrote:
I just had a look at the chart showing GB Women's Rank Group with Time. Instead of using a single icon type at each data point, might it be useful to show either a directional arrow (up/down) or a solid dot, to indicate how a player has progressed (in rank) from the previous rankings set? Apologies, if this has already been suggested/discounted.

First, welcome foobarbaz

Second, that's an interesting idea. I think I know generally what you mean. But, I'm not sure exactly which chart you're proposing to implement it on? Which page of this thread was it on?

AliBlahBlah · All-time great

indiana wrote:
AliBlahBlah wrote:
Still trying to figure out CH - adding just a note seemed to dissociate it form it's position in the chart, I couldn't 'see' where CH fell just from reading it and trying to imagine it.

[ ... ]
I tend to disagree and think it a pretty good mix of the graphical and tabular. I can see folk looking at the graph, which I do really like as is, and the note of the CH just to the side. It doesn't have to be actually contained within the graph to be associated.
If you find your elegant solution good but personally my eyes don't take to too much in a graph, call it clutter or whatever. I like to see and get it fairly quickly.

It's just information, and unrelated to the plot area, and so it shouldn't be there, unless it's in context.
Here:

Putting that number there doesn't add anything to the chart - you can't really visualise where any of the CH values lie accurately, and so they don't convey how far or close current the data plotted in the bars and lines is to that value. It's just there because it's technically possible to put it there, and we know those values.
I need a third axis - I'm already using a second axis to plot the two lines. Then I could just mark a small discreet yellow dot in place. This should be possible. I just don't know how.

It brings up a thorny question, as to chart design, with which I struggle greatly.
e.g.

I like this:

But, this conveys more information, but less concisely. Which is'better'?

Or, these two charts essentially show the same information. Which is better?:

I prefer the top one.

But, I never know what to do in any given case, and spend hours trying things out unsuccessfully.
e.g. it took me fully 6 hours to figure out how to sort the top of those last two charts so that ech GB rank was sorted in order of incumbency, most to least weeks. 6 hours!

indiana · Tennis legend

Re our discussion about how to add CH information to rankings movement over the year, as a consumer your illustrative CH in brackets to the side works for me. It adds to the whole thing and for me doesn't have to be within the graph itself. Yes, such as a small yellow dot may be a non cluttering alternative that I sense is probably preferable to you but I think I'd still be at least as happy with the CH to the side. The range for the year and the CH is all there.

Re your other chart designs, interesting and I will try to take time to consider further and give a consumer response re these.

Michael D · All-time great

On the most frequent incumbent, ABB, I prefer the graph at 3.0, rather than the colourful shapes in 3.1 The latter has some strengths eg I can see at a quick glance that Sam Murray only featured in the #10 position, but overall the graph is clearer and easier to digest.

The wavy colourful lines above this I find pretty indecipherable, since I can't absorb the key and the lines at the same time, which one needs to be able to do to know who is who in the wavy lines. The top one is easier to look at because it is stiller, but the colours are far too near each other especially the various blue green shades which at this relatively small scale I have huge difficulty in distinguishing between.

AliBlahBlah · All-time great

Michael D wrote:
[...] but the colours are far too near each other especially the various blue green shades which at this relatively small scale I have huge difficulty in distinguishing between.

13 distinct shades that appear visually unique is hard. I used a colour wheel to optimise it, too. The 2018 only version works better, as less players are involved. I messed around assigning different players different colours, but there's too much movement between ranks, eventually something meets and blends into something similar to it. Adding a grey or black or white to it looked awful for some indefinable reason.

I like the idea and compactness of the layout. It just needs to be used judiciously - with fewer series, say maximum of 8.

foobarbaz · Strong Club Player

AliBlahBlah, thanks for the welcome smile

The chart doesn't appear to be on any of these pages, I (wrongly) assumed it had originated here. The link to it is https://public.tableau.com/profile/insomniacfolder2933#!/vizhome/2017-18PlayersinEachGBRankovertime/Sheet1 . I chose rank group 4, and it seemed to me that when one player is succeeded by another in the rank group it's not clear whether the latter player is on the rise, or sliding down from higher rank groups. It's only really the first data point in a players rank group 'sequence' that could do with the extra clarification provided by a directional icon, but then visually it might work for all the data points?

indiana · Tennis legend

Re 2 & 3 above, I'll pass on 2, I find both rather difficult to generally take in with any ease though here not just because of the distinction between colours that Michael refers to but more generally. And particular weeks at particular rankings doesn't grab my interest anyway, comparative to the overall number of weeks of 3. It may be something that is difficult to give an appealing overall picture to and the more simple totals of 3 are much easier to clearly illustrate.

Re 3, I prefer 3.1. A bit different and very easy to get a quicker grip of which players have been most at each rank and additionally almost instantly over which rankings any particular player has featured, eg Sam just #10 in the weeks that she has been top 10, Naomi very much most of the year at #3 with a few weeks at each of #2 and #4. The only thing would be that blue to green colour distinction between such as Gabi and Freya ( while appreciating the probable difficulties of 13 different shades that one for instance seems very close ). The names in the blocks though sorts out some difficulties and overall I do certainly like 3.1 better though 3.0 gives the absolute numbers for all if someone wants that.

AliBlahBlah · All-time great

foobarbaz wrote:
AliBlahBlah, thanks for the welcome
The chart doesn't appear to be on any of these pages, I (wrongly) assumed it had originated here. The link to it is https://public.tableau.com/profile/insomniacfolder2933#!/vizhome/2017-18PlayersinEachGBRankovertime/Sheet1 . I chose rank group 4, and it seemed to me that when one player is succeeded by another in the rank group it's not clear whether the latter player is on the rise, or sliding down from higher rank groups. It's only really the first data point in a players rank group 'sequence' that could do with the extra clarification provided by a directional icon, but then visually it might work for all the data points?

Hmm.. OK, I see what you mean. That is a lot more complicated than it seems. I'm pretty sure I could add something to the tooltips, but actually putting it in the plot area is harder.

It's not origianlly intended as a slice across the players journey up and down the GB ranks, but across the GB rank itself, thus, as originally formulated, where the player came from, or indeed, where they go to, wasn't so important. The thing was to look at the fluctuations within each position.

It would be useful ancillary information. I don't have any plans to keep that file up to date over the long term though.

AliBlahBlah · All-time great

indiana wrote:
Re 2 & 3 above, I'll pass on 2, I find both rather difficult to generally take in with any ease though here not just because of the distinction between colours that Michael refers to but more generally. And particular weeks at particular rankings doesn't grab my interest anyway, comparative to the overall number of weeks of 3. It may be something that is difficult to give an appealing overall picture to and the more simple totals of 3 are much easier to clearly illustrate.
Re 3, I prefer 3.1. A bit different and very easy to get a quicker grip of which players have been most at each rank and additionally almost instantly over which rankings any particular player has featured, eg Sam just #10 in the weeks that she has been top 10, Naomi very much most of the year at #3 with a few weeks at each of #2 and #4. The only thing would be that blue to green colour distinction between such as Gabi and Freya ( while appreciating the probable difficulties of 13 different shades that one for instance seems very close ). The names in the blocks though sorts out some difficulties and overall I do certainly like 3.1 better though 3.0 gives the absolute numbers for all if someone wants that.

Thanks to you and others for the feedback. It's all useful.

Generally, this is almost a rhetorical question. There are at least a dozen ways to present data in any situation, and dozens of viable variations and layouts again of each of those. Finding the best, most clear, concise, and meaningful arrangement is an impossible - how long is a piece of string type - endeavour; at least for me. Invariably, I just end up frustrated, and go with 'the last thing I have that wasn't awful'. I've only ever gotten one chart to do exactly everything I set out to do with it.

foobarbaz · Strong Club Player

Yes, I can see it's not a trivial exercise to add this extra nuance to each of the series in each of the GB rank group plots. First there is the capturing of direction of travel in the WTA ranking from week to week for each player. Then the use of that variable to modify (potentially) each data point icon in the plot series for a player, within the overall GB rank group comparison dataset.

I started tinkering a bit with python/matplotlib, and while it's possible to change icon colour within a plot series, icon types are more tricky. I saw a solution where the different icon types are plotted as separate groups, but then the player data series are lost (I think). I wonder whether a simple solution would be to just to overplot with both down/up series of 'icon modifier' symbols for the entire GB rank group as one, but then this might be like trying to combine two datasets into one graph ...

I can see this is going beyond the original intention of the chart in question, and hear what you say about the charts long term upkeep.

Is the wta ranking dataset freely available, it'd be nice to keep pondering the above with the actual data, rather than trying to create a mock-up dataset?

christ · Futures level

It's like a whole 'nother language.