Author Topic: Visualizing the OYTT  (Read 137839 times)

red marley

Visualizing the OYTT
« on: January 08, 2015, 08:04:58 pm »
Edit: You can view the OYTT visualization app here: http://gicentre.org/oytt



Here are a couple of screenshots from my attempt to visualize the progress of all three riders in the One Year Time Trial (click to enlarge).



Godwin's 1939 progress shown as the grey line, Steve in red, IronOx in green, and when he appears, Tarzan will be in blue.

This is basically a time against cumulative distance graph, but standardised with respect to the distance covered by a ghost Tommy Goodwin continuously riding at his average pace of 205 miles per day. So a rider who managed 205 miles per day would progress horizontally - any faster than that and they would move upwards, slower and they start sloping down. A stationary rider would slope down parallel to the steeply sloping lines seen towards the right (you can spot Tommy's only rest day on October 28th).

The vertical axis therefore shows how far ahead or behind Tommy's ghost a rider is at any point in time. As you can see, Tommy himself was below his 205 miles per day speed between January and May, after which he started to increase his daily pace until November. By standardising distances in this way, it should be possible to compare all four riders' progress even though they started at different times.

The brownish zones to the right represent the previous world records between 1911 and 1939, so if all goes to plan, we should see those being hit from around June onwards.

After the first week, you may just be able to make out that Steve (in red) is above Tommy's actual Jan 1-7 pace, but below that year-long average.  IronOx, with only short rides so far, is already 700 miles behind Tommy's ghost. Below is an enlarged section showing week 1.



This shows food and sleep stops (downward sloping lines). Should be interesting to compare once Tarzan starts and we get to see the tortoise and hare together.

I am still sorting out a few minor bugs in the interactive version of the software (damn you IOS), which will allow you zoom and pan around the charts, but it should appear here soon. The software is being fed by the GPS data from Strava, so should get updated once per day.

Any comments on the design / suggestions for improvements are welcome.

Re: Visualizing the OYTT
« Reply #1 on: January 08, 2015, 08:21:41 pm »
That's clear enough :thumbsup:
DJR is my 27 year old bike, custom built by Dave Russell & heavily refurbished in 2009. Also, a carbon Beone parts bin special built in 2011 and no longer have Brompton after its road rage demise!

Kim

  • An appetite for the epic, but no real stamina
Re: Visualizing the OYTT
« Reply #2 on: January 08, 2015, 08:22:36 pm »
Yeah, that's really good.   :thumbsup:
Watching the TV without subtitles is like riding up a hill without using the gears :)

Re: Visualizing the OYTT
« Reply #3 on: January 08, 2015, 08:28:13 pm »
I really like it, particularly the time and cumulative distance. It brings an immediacy to it while illustrating the enormity of the task.

I am sure Steve himself would find it motivating.

Wowbagger

  • Dez's butler
    • Musings of a Gentleman Cyclist
Re: Visualizing the OYTT
« Reply #4 on: January 08, 2015, 08:28:31 pm »
Clear enough for me with my not-fantastic eyesight, but it didn't seem to enlarge when I clicked.
You won. Get over it.

simonp

  • Omnomnomnipotent.
Re: Visualizing the OYTT
« Reply #5 on: January 08, 2015, 08:29:38 pm »
A very nice visualisation. If I’m interpreting this correctly, it shows also that Tommy Godwin’s pace dropped further in mid January and he didn’t reach the average pace until the middle of May. If Steve keeps up his current pace he could be 1500 miles ahead by the end of February.

I notice the way you’ve shown IronOx starting on 5th January. ISTM that this could be confusing as his start date is not 1st January, so the x axis isn’t normalised. Would it better to time shift IronOx and Tarzan back to allow for their delayed starts?


red marley

Re: Visualizing the OYTT
« Reply #6 on: January 08, 2015, 08:43:53 pm »
I notice the way you’ve shown IronOx starting on 5th January. ISTM that this could be confusing as his start date is not 1st January, so the x axis isn’t normalised. Would it better to time shift IronOx and Tarzan back to allow for their delayed starts?

I think of it as a 'virtual partner' ghostly Goodwin setting off at the same time as each of this year's riders. The vertical position of each of the lines therefore represents the distance between the rider and their virtual Godwin. So in that respect they are 'normalised'. I wanted to keep the absolute times for each entry correct so displacing some of the lines back to Jan 1st would make it a little misleading in my opinion. The only aspect of the chart that is not consistent for all riders, is the horizontal position of the world record bands. These all assume a start date of Jan 1st, so for IronOx and Tarzan, those steep lines would each shift to the right by 4 and 10 days respectively.

LittleWheelsandBig

  • Whimsy Rider
Re: Visualizing the OYTT
« Reply #7 on: January 08, 2015, 08:46:54 pm »
Tommy had lots of falls on ice early on, so his mileage was adversely affected.
Wheel meet again, don't know where, don't know when...

Re: Visualizing the OYTT
« Reply #8 on: January 08, 2015, 08:48:15 pm »
I notice the way you’ve shown IronOx starting on 5th January. ISTM that this could be confusing as his start date is not 1st January, so the x axis isn’t normalised. Would it better to time shift IronOx and Tarzan back to allow for their delayed starts?

I think of it as a 'virtual partner' ghostly Goodwin setting off at the same time as each of this year's riders. The vertical position of each of the lines therefore represents the distance between the rider and their virtual Godwin. So in that respect they are 'normalised'. I wanted to keep the absolute times for each entry correct so displacing some of the lines back to Jan 1st would make it a little misleading in my opinion. The only aspect of the chart that is not consistent for all riders, is the horizontal position of the world record bands. These all assume a start date of Jan 1st, so for IronOx and Tarzan, those steep lines would each shift to the right by 4 and 10 days respectively.

Jo,

That works for normalising against the average of Tommy (the X axis) however it makes comparing against Tommy's actual ride and between riders harder as you have to mentally shift the later starters.

red marley

Re: Visualizing the OYTT
« Reply #9 on: January 08, 2015, 08:54:28 pm »
Fair enough. I guess the value of absolute vs standardised start times depends in part whether it is length of time riding, or time of year / events (e.g. PBP) that has a greater effect on progress. I will add the ability to switch between the two in the interactive version.

Re: Visualizing the OYTT
« Reply #10 on: January 08, 2015, 09:42:36 pm »
I must be misunderstanding this, but I would have thought that the area under the negative part of the curve for Tommy Godwin would match that of the positive part, that is, an equal amount riding below pace as above it, to reach the average?
“That slope may look insignificant, but it's going to be my destiny" - Fitzcarraldo

red marley

Re: Visualizing the OYTT
« Reply #11 on: January 08, 2015, 09:53:28 pm »
The speed riding when below the line is, by definition, slower than that above the line, so it contributes less to the total distance (and therefore annual average) than speedy riding. To take an extreme example, if a rider was stationary for the first 6 months then rode at some fixed speed for the second 6 months, the entire graph would be under the 0 line.

It's like the effect on average speed of riding up and down hills. Yes, the shooting down hills raises your average speed, but not as much as slogging upwards lowers it as you tend to spend more time climbing than descending.

[Edit] Or to put it another way: The zero line represents the passage of the Ghost of Godwin who rides continuously at 8.5 mph (=205 miles per day). The time when under the line represents the period when the ghost is ahead and the time above it when the rider is ahead. These don't have to be equal - Ghost of Godwin could be ahead for almost the entire year, only to be pipped at the post on the 31st of December (and please Steve, don't do that as we'll all be having heart attacks over Christmas).

simonp

  • Omnomnomnipotent.
Re: Visualizing the OYTT
« Reply #12 on: January 08, 2015, 10:16:33 pm »
In fact it’s entirely possible for all but the last point and the first to be below the line. As it would have been the case if the attempt had ended when it crosses the line from below to above in August.

Re: Visualizing the OYTT
« Reply #13 on: January 08, 2015, 10:30:33 pm »
I must be misunderstanding this, but I would have thought that the area under the negative part of the curve for Tommy Godwin would match that of the positive part, that is, an equal amount riding below pace as above it, to reach the average?

If, instead of cumulative distance, the graph had plotted daily distance (or weekly, or any other periodization), then you'd be right: the areas above and below the average would be equal.

Re: Visualizing the OYTT
« Reply #14 on: January 08, 2015, 10:47:49 pm »
I must be misunderstanding this, but I would have thought that the area under the negative part of the curve for Tommy Godwin would match that of the positive part, that is, an equal amount riding below pace as above it, to reach the average?

If, instead of cumulative distance, the graph had plotted daily distance (or weekly, or any other periodization), then you'd be right: the areas above and below the average would be equal.

That can't be right: it plots - or will - annual distance. No?

Re: Visualizing the OYTT
« Reply #15 on: January 08, 2015, 10:55:46 pm »
Got it thanks, and it also helps explain why the previous record lines aren't vertical.
“That slope may look insignificant, but it's going to be my destiny" - Fitzcarraldo

simonp

  • Omnomnomnipotent.
Re: Visualizing the OYTT
« Reply #16 on: January 08, 2015, 11:05:57 pm »
I must be misunderstanding this, but I would have thought that the area under the negative part of the curve for Tommy Godwin would match that of the positive part, that is, an equal amount riding below pace as above it, to reach the average?

If, instead of cumulative distance, the graph had plotted daily distance (or weekly, or any other periodization), then you'd be right: the areas above and below the average would be equal.

Gareth is describing a different visualisation, still spanning a year, breaking the year up into equal-time segments, and simply plotting the distance covered in each segment.


That can't be right: it plots - or will - annual distance. No?

Re: Visualizing the OYTT
« Reply #17 on: January 08, 2015, 11:55:33 pm »
I must be misunderstanding this, but I would have thought that the area under the negative part of the curve for Tommy Godwin would match that of the positive part, that is, an equal amount riding below pace as above it, to reach the average?

If, instead of cumulative distance, the graph had plotted daily distance (or weekly, or any other periodization), then you'd be right: the areas above and below the average would be equal.
That can't be right: it plots - or will - annual distance. No?
Gareth is describing a different visualisation, still spanning a year, breaking the year up into equal-time segments, and simply plotting the distance covered in each segment.

Fair enough.

I think I was misled by the fact1 the current graph doesn't plot cumulative distance, in fact doesn't plot absolute distance at all - it merely shows cumulative deficit/surplus vs the ghosted average Tommy.


1: Coupled with a lack of thinking it through

red marley

Re: Visualizing the OYTT
« Reply #18 on: January 09, 2015, 07:25:33 am »
First 8 days relative to start date:



Steve continues to steadily improve on Godwin's January milage with a remarkably consistent performance.

Re: Visualizing the OYTT
« Reply #19 on: January 09, 2015, 08:06:09 am »
Thank you Jo   This makes it all much clearer.
Only those that dare to go too far, know how far they can go.   T S Elliot

Jack_P (Burlycross)

  • ton of tons for 2016
    • Burlycross riding is life
Re: Visualizing the OYTT
« Reply #20 on: January 09, 2015, 08:20:26 am »
That's superb, really shows how Tommy Godwin had a Stella Summer riding period.

ianrauk

  • Tattooed Beat Messiah
Re: Visualizing the OYTT
« Reply #21 on: January 09, 2015, 10:29:25 am »
Thanks Jo.
The graphs really help visualise Steve's challenge.

Mr Larrington

  • A bit ov a lyv wyr by slof standirds
  • Custard Wallah
    • Mr Larrington's Automatic Diary
Re: Visualizing the OYTT
« Reply #22 on: January 09, 2015, 01:00:13 pm »
Top SCIENCE, jo :thumbsup:
External Transparent Wall Inspection Operative & Mayor of Mortagne-au-Perche
Satisfying the Bloodlust of the Masses in Peacetime

red marley

Re: Visualizing the OYTT
« Reply #23 on: January 10, 2015, 12:05:17 am »
I've added both Steve's (red dashes) and Tarzan's (blue dashes) planned schedule for the year. Steve has provided two schedules depending on how things go (see his site for rationale) so I've shaded the area between them.



I hadn't realised quite how ambitious his target was compared to Tommy Godwin's and even Tarzan's. That 10 day tactical advantage that Tarzan has isn't going to count for much if they both stick to their schedules as Steve would be between 4,000 and 8,000 miles up in the last two weeks of their rides.

Over the first 9 days, Steve is on target for his more optimistic schedule  :thumbsup:



Should get interesting once Tarzan starts tomorrow.

Re: Visualizing the OYTT
« Reply #24 on: January 10, 2015, 12:10:44 am »
Thoroughly fascinated by your "visualisations" Jo.

But I don't know you, and I have a Jo in my team who's a JavaScripter. You aren't the same Jo are you?
You're only as successful as your last 1200...