Discussion: Rogaine Network Map

Righto - so was teaching myself some network map visualisation stuff this weekend so decided to have a crack at doing one (or a few) on WARA.

So after painfully scraping 150 or so results pages and manually transforming the data I came away with some cool graphs



So what do these show - well I did 2 sets of 2 graphs - 1 set shows full linking which means if oyu had a partner you have a link to them and from them and the other set is just 1 link which is a straight line between you and your partner(s)

Within those graphs I did the full network every person who has rogained and a cut down version for people who have done over 10 events and have ahd multiple partners.

Links below:
Full network - full linking: http://glimmer23.com/rogaine/index.html#rogaine1.g...

Full network - single linking: http://glimmer23.com/rogaine/index.html#rogaine2.g...

Rogainers with more than 10 races and multiple partners - full linking: http://glimmer23.com/rogaine/index.html#rogaine3.g...

Rogainers with more than 10 races and multiple partners - single linking: http://glimmer23.com/rogaine/index.html#rogaine4.g...

Give them time to load they are big ass files (well the first 2 are).

Also yes there are data quality issues - e.g. duplicate names, misspelling, name changes etc etc I wasnt about to go clean scrapped data for a training exercise - but if WARA wants it done properly I will happily use their DB to do it right!

Other interesting insights:
there has been ~38,000 participants in WARA (~10,500 Unique)
Tooms and Sarah Dunnage hold the highest score record of 5,460
Bryan Mcclintock has the most events at 93
Tricky has highest points collected at 225440 in 91 events
Highest average is 3,900 but thats 1 event - so lets play fair and say 10 events or more - Tim Skima - 2,935
and the sad stat 70% of competitors only do 1 or 2 events.

Oh and there is 6 color grouping on the chart cold to hottest based on events I think its something like

blue 1-2
turquoise 3-4
green 5-7
yellow 8-12
orange 13-19
red 20+

Or something like that - you get the drift.

as it is not obvious - zoom right in to see all the peoples names or use the search function

There's at least three names on my graph that I have no idea about. Who the devil is Dan Mullan? :)

And what's the difference between an inbound and outbound link?

Nothing on the main mask because if you rogained with someone I gave person 1 to person 2 and person 2 to person 1 a link

Rather than just 1 to 2. - so inbound and outbound on the full linking graphs 1 and 3 will be matched.

Also Dan Mullan you did the novelty with 2 years ago with Dave and I. :) HUZZAR ahah

Do you count events where a 'withdrawn' was recorded? I'd have several more if that is the case. Actually I clicked on my dot (yes I found it) and one guy I competed with once and had a 'withdrawn' result is listed so I guess it does.

I was surprised to find I'd competed with Carole Baxter, Ron Lockley and Geoff Monk but then I believe that's an event where I got injured beforehand, volunteered myself but went out and competed with them for a couple of hours anyway :-)

Also I'm fairly sure some people (e.g. MrRogaine) have reached 100 events. I'm disappointed that MacTickTock has more events than me given I started many years earlier but then I've missed several recently and he still competes when he volunteers (truck driver) whereas I don't.

yes well I only had access to the rogaine results as per the website which only go back to 1996 - I am sure there are plenty before then.

Yes if you are on the sheet you will be one the results - had to get rid of some duplicated results but by large if its on the 7 pages of results listing (except the 2 bike rogaines) youll be on the chart.

btw - there is a search function at the top you didnt need to zoom in and search for your name - just type it in and it will show you you - also if you click on anyones names on inbound/outbound it will change to them.

Ahh okay that explains it. I had at least another nine events pre-1996 so I guess I've made the 100 club by now too.

Haha, Nick Lethbridge - 90 events, 2 different teammates.

Next stat request - who's the biggest rogaine tart? (highest teammate/event ratio)

ok so highest average teammates - the tart award. I think I can work that out tonight shouldn't take more than a few minutes.

If you take it from one given year then I might win. I think I had six teammates in six events with six top tens one year.

You may have to do it as >10 events too otherwise you'll get people with two events and four different people or something like that.

What's the difference between an inbound and outbound link?

Oh and there is 6 color grouping on the chart cold to hottest based on events I think its something like

Must be based on something else. I've done one event with Andrew Knight and that's blue whereas I've also done just one event with Richard and that's green. Jen's link is also blue (one event), Mark Dowley is yellow (definitely not 8-12, more like 3-4) and I have a whole stack that are orange but for whom I would not have done 13-19.

Do you mean that's how many events the other person has done?

that the other person has done - not done with you.

The node coloring relates to the node (person themselves - not the person whom they are linking).

re inbound vs outbound - in scenario 1,3 nothing - its matched - see my earlier response to Paul. in scenario 2, 4 I got rid of the dual linking and made it 1 direction so inbound and outbound isnt the true sense what its supposed to be used for but just shows connections in total.

Where it works inbound and outbound is like phone calls - these people rang me and I rang these people. doesnt really work for the analysis I have done.

No, it doesn't - you call me all the damn time, but generally with unsuccessful attempts at smacktalk. <3

I did a rogaine with Bob McCreddin?

Lots of ones missing from my record, how far back in time did you go?

1996. Geez even I can read (some of) the comments.

re inbound vs outbound - in scenario 1,3 nothing - its matched - see my earlier response to Paul. in scenario 2, 4 I got rid of the dual linking and made it 1 direction so inbound and outbound isnt the true sense what its supposed to be used for but just shows connections in total.

This makes even less sense now than before you explained it.

Where it works inbound and outbound is like phone calls - these people rang me and I rang these people. doesnt really work for the analysis I have done.

I'm sure I emailed some of my teammates. Would 'outbound' mean I phoned (or emailed) them more than they phoned me? If that was the case, I should have 100% outbound.

I am outbound to Shenton College to check access for the Nav Dash in April . When I am next inbound I will be able to read the next 50 comments

Richard I can save you some time. Given it's on the same weekend as the ultra-long, there's no point even trying because I won't be there.

I have had 13 partners in 13 events, so I guess that's somewhat tarty. I only wonder why I can't find myself in the "more than 10 events" graphs.

ok let me try again

Inbound links means someone connecting to you
Outbound links is you connecting to someone else.

In graph 1 and 3 - you will have dual linking (both in and out) as I created the links as such
Ricky (id 1) rogained with Paul (id 2)
Links created:
1 > 2
2 > 1

Thus you will both have each other on inbound and outbound.

In an attempt to reduce the amount of links I wanted to see only the truly unique links. - in graphs 2 and 4 I did this by saying (in the example above) 1 > 2 and 2 > 1 are effectively the same so just keep one of them - 1 > 2 remained.

In this scenario Ricky would be outbound to Paul and Paul would have inbound from Ricky. but not the corresponding side of the transaction.

If this doesnt make sense I suggest you ignore inbound and outbound :)

sten likely because the 13 people you did it with didnt continue to rogaine :) thus when i filtered the must have 10 events and must have multiple connections - all of your connections dropped out because they hadnt done 10 events :)

Cool - love seeing the event and score sum, and clicking through the past teammates.

Can the graphing program show a thicker line for more events between a couple of people?

Can you show a graph with number of events before having kids and number after? I want to see how many people drop out in that scenario.

Actually one of my partners has also done more than 10 events, but I guess we fail the "multiple connections" criterion.

Most tarty, by the way, would be many one timers with four partners. Some who I have rogained with have a partners/event ratio of 4 that way.

Andre - technically yes, in fact it can also show bigger nodes for more connections or on any variable I wish etc. but I am still learning so havent explored these yet - first time using the program yesterday and it was crashing something fierce.

Yes I suspect it will be a family of 6 who may be the tartiest.

Bryce, if i understand correctly (very unlikely!) the graphs are showing me as having done 14 events. I have done way more than that so I guess you have not gone back to day 1 of WARA?

When does the data start?

http://www.attackpoint.org/discussionthread.jsp/me...

Maybe it's only counting your top ten results.

Don't you only do about one event per year?

yes well I only had access to the rogaine results as per the website which only go back to 1996 - I am sure there are plenty before then.

Sigh. Read more thoroughly Richard. Or look at the screen around your glass of wine instead of through!

Well there is a job for some lucky volunteer to go back to the beginning of time in 1980 - the results are in the Newsletter archive

I did two Rogaines last year.

Plus I already answered his question immediately after the last time he asked it!

I don't think you've done that many since 1996 have you because I always knew you as an orienteer? You've probably only done 14 in the last 20 years, which is why someone was aghast that you could get a life member award.

Also I think it would have taken long enough for Bryce to copy and paste the names from the website, let alone typing them all in manually from the newsletters.

yes transposing the information off the 150 pages was a few hours effort - manipulating them in excel to the format I needed was another few hours.

from that point playing with the data and getting it into the program was painful but easier.

re going pre 1996 - surely they must be stored on pebbles? if I took all the pebbles files I could do more - the bigger challenge you face is there is probably a large number of incorrect spellings/double entries etc in there there would be a good day or two work in data remediation alone.

It would be nice though when WARA gets to the point of better inter connectivity of users and its website to have something like this available.

Also the real power of things like this is you can actually see who the people are who bring in members and who are people who have been around a while - if you wanted to increase member retention and growth you would focus efforts through these people - but thats opening a whole different strategy can of worms which I dont really intend on entertaining atm - this is more an interesting picture of history

Pebbles used to be called Fred. I'd be surprised if electronic results existed in the early days but it did at least in 1993.

I won the WRC in 1998!!!!! would have been a good time to retire

Nice little archive of WRC results that could make another Bryce Data Set

Two more rogaines in Arizona while I was living in Canada

That must count for something

I believe you won the mixed category, not the whole thing. Geez you didn't even get a top ten badge.

A win is a win and I have the medal to prove it

In that case, I won the ARWC in Tasmania! I have the cheese board to prove it (somewhere).

Richard - not sure how interesting ti would be as I doubt there would be that much team mate swapping - where as in WARA there seems to be a lot of ti - WRC maybe not so much.

there is heaps of potential analysis wara could do - like what format attracts the most people, is there a bunch of setters that set events which are of higher popularity...

more importantly how do we convert the 70% of once off or twice to something much smaller. you really want people coming back and bringing their friends. I was actually surprised over 10 years of data that there wasnt more people with 10+ events and multiple links.

People always come to events set/vet by me.

Crunchy Nut Clusters - 387
The Dark Swamp - 417
Karn't Breakout - 275 (that's high for a 6hr)
Metrological - 283 (BA was involved so that scared people away)

I probably set some others before that but they're not important.

The novelty event attracts the most people. They should hold more of them on Saturday afternoons throughout November to February. I doubt there'd be anything similar on during that time.

I have just done a count I have done 12 Oz Champs since 1982, worst performance was with Craig Dufty 1993 13th Forest. 9 different partners, most is 3 with with Paul Hoopmann

Not surprised, Craig is pretty terrible at this sort of thing.

Selina Wilson has rogained with 56 partners with an honorable mention to MrRogaine with 48. However they aren't the tarts because the've gone a lot of rogaines in fact Selina is in second place with 92.

Now as someone rightly pointed out - if you did 1 rogaine in a team of 6 - you would be the tart along with anyone in that team who only did one - so we are only looking for people with 10 rogaines or more.

Well after playing with results I think 9 was a better statistical outlier - Alex Craven with 2.44 per rogaine. but given the entire population is 1.73 we can say generally not that tarty. People arent sharing their partners much.

If you rogained in a team of six then you should be publicly flogged for breaking some rule or other.

Think biggest team I found was 7.

DSQ

R1. A team shall consist of two, three, four or five members.

But when you're with another person tRicky I'd hardly call it a team effort... *stir, stir*

Does a female team have any members?

I suggest Selina Wilson can be WARA's Kevin Bacon. I guess most of you would be within six degrees from Selina, because even I am.

Sten -> Kathy Haskard -> Ann Smithson -> Phil Dufty -> Rob West -> Moreen Cox -> Selina Wilson

yeah I got same number of hops as you via rob or lois west

16 rogaines on the list. I did at least 3 that pre-date it, which would also add another 3 partners. Plus a bike rogaine (another partner). But a fair whack of the partners there are school kids from the two rogaines I did leading school groups around a decade ago. We let them get lost on their own these days...

You cheated on that bike rogaine so it doesn't count.

Surely it still counts as doing the rogaine? Unfortunately, I have lost my prize (Great Truckin' Hits CD, featuring classics such as "Give me 40 Acres (and I'll turn this rig around), 15gears and 14 wheels and My Diesel Friend), so I can't return it :(

Well if I get my rogaining stats from Jim Langford, he doesn't count events in which you are listed as 'withdrawn', which includes the event where Bryan and I were out for the whole 24hrs with enough points for 2nd place but 'accidentally separated' near the finish.

So yeah, I'd count it.

As to the 'six degrees', well I get to Phil via Jim Langford so I'm a lot closer.

You can do one better, tRicky:

tRicky -> Vince Harding -> Warren Smith -> Kerry Cowie -> Selina Wilson

I don't have time to search through all that. Oh wait, yes I do. What I meant is I don't have the patience.

It's entirely possible I may get even closer. I know that Kerry Cowie and Samantha Robinson were somehow involved but Sam only has a link to two people at the moment because she stopped rogaining after I took her out a couple of times (as most people tend to do) in 1996 so she may have a link with Kerry prior to that.

I could also get to Warren via Geoff Monk, Ron Lockley or Sue Joyce. Rogaining is so inbred.

What, most people take her out a couple of times, or should you have inserted the bracketed text between 'rogaining' and 'after'?

Zing!

No. Most people stop rogaining after I take them out a couple of times.

I've never been on date with tRicky and I've stopped rogaining. I'm not sure that Ricky and his date selection is the sole reason keeping participant numbers low.

What??? You mean... tRicky dated all those girls?? And they said No???

I don't want to even think about tRicky's date.