Cameras + YouTube + Google - using the power

[BentMikey]

Whilst repeat offenders are fairly small, I've little doubt that that will go up in future with the ongoing explosion in the number of cyclists using cameras (and other road users for that matter).

I wonder whether the seriousness of videos is something worth looking at? Maybe more along the lines of what an editorial group rates videos as, rather than just the video owner, and that a set of tags gets generated as a result.

 

[BSRU]

So, to return to the original purpose of the thread. I think the potential that this combination of technologies provides is to improve the signal-to-noise ratio when using the available video clips for some particular purpose amongst the many different reasons that people had for posting them.

One such purpose is highlighting bad driving in an attempt to get some kind of change that might eventually improve the lot of cyclists on the road. There seem to be (at least) two potential ways to select videos for this purpose: repeat offenders (or at least vehicles that appear in more than one video), and selection based on the seriousness of the incident.

The "repeat offenders" one is probably a non-starter given the relatively small number of cyclists who post videos. I spent a bit of time playing with the YouTube API and wrote a program that extracts registration numbers from the titles of all videos on a particular channel, does a search of YouTube for any other videos with the same registration number in the title, and then lists them under the registration number. I was not surprised to see that there are very few repeat offenders turned up, even for some of the London commuters' channels - Mikey and Gaz for example. (I had fun writing it, though.)

So that leaves the seriousness of the incident as a basis for selection. There are many videos that, whilst showing behaviour that undoubtedly causes increased risk to the cyclist, are subject to debate even among cyclists as to whether the behaviour was sufficiently dangerous to warrant highlighting. (You only need to look at this topic for proof of that assertion.) It seems to me that the best approach, then, is to "pick the low hanging fruit" and concentrate on videos showing incidents that leave little or nothing open to argument regarding whether the driver's behaviour was dangerous. (I know, there will always be some 'ard man who claims that a truck screaming past at 40mph three inches from their elbow is nothing to make a fuss of, but I think we can just ignore them.)

Earlier, I suggested that some standard for tagging videos might be useful to aid in this, and I do think that might serve a purpose if enough people starting tagging according to the standard. However, it seems like that may be a sledgehammer to crack a nut; I've been involved in standards-making in the past and getting agreement on even the most trivial detail can be a right pain.

So, for those that are interested, why don't we just each share links to a small number of videos that we think are extreme examples and make a page of them? No fancy searches, just people. If anyone doesn't want to join in, then they are quite free to remain silent.

Do you mean have a dedicated YouTube channel managed by someone, where videos which are deemed worthy by a group of reviewers are then favourited on that channel?

 

[beastie]

Why is it that cyclists who claim to be god, really are just sanctimonious pricks with an odious personality, who show zero respect for anyone.

Not very god like.

 

[Origamist]

So that leaves the seriousness of the incident as a basis for selection. There are many videos that, whilst showing behaviour that undoubtedly causes increased risk to the cyclist, are subject to debate even among cyclists as to whether the behaviour was sufficiently dangerous to warrant highlighting. (You only need to look at this topic for proof of that assertion.) It seems to me that the best approach, then, is to "pick the low hanging fruit" and concentrate on videos showing incidents that leave little or nothing open to argument regarding whether the driver's behaviour was dangerous. (I know, there will always be some 'ard man who claims that a truck screaming past at 40mph three inches from their elbow is nothing to make a fuss of, but I think we can just ignore them.)

When I was putting together my defensive cycling videos (yuk - I sound like such a media darling), I used DfT and TFL cycle collision data as the basis for this series. This meant that the severity of the incident was not the criterion I adopted (although it is clearly entangled). Instead, I wanted to illustrate the most common types of collision configuration that involved cyclists getting injured and attendantly highlight approaches to anticipate or mitigate such situations.

I'm not sure this helps greatly, but it does give a different perspective on the basis for selecting seqeunces. My limited approach has obvious drawbacks though.

 

[BentMikey]

I was hoping you'd post a bit on this, Origamist, as your approach on reporting offenders is very different. I can see some of the advantages to it, but also disadvantages just as with the public approach I follow.

 

[Origamist]

I was hoping you'd post a bit on this, Origamist, as your approach on reporting offenders is very different. I can see some of the advantages to it, but also disadvantages just as with the public approach I follow.

Yes, for those that don't know, I have two accounts. One for general, broadly educational vidoes (the defensive cycling series) on YouTube and a different account on Vimeo for reporting grade "A" idiots.

The former account does not list reg nos in the title, description or tags, but the reg plate is often visible in the film. Effectively, this means the chance of the driver finding the video on YouTube is fantastically remote. This channel is accessible to anyone with too much internet time on their hands and I do not use it for naming and shaming. Although occasionally in the summary section I might comment as to why I think it was poor driving, cycling, motorcycling etc. Comments are rarely disabled.

If I report someone to the Met Police's RoadSafe initiative, I use a private Vimeo account (different user name) and I only send the password to the police. I disable comments on the site and I rarely use this account.

I like to keep the two accounts separate - for various reasons.

 

[Little yellow Brompton]

Why is it that cyclists who claim to be god, really are just sanctimonious pricks with an odious personality, who show zero respect for anyone.

Not very god like.

Jehovah, Zeus, Odin, Jupiter ? You description ( above) seems to fit them all...

 

[classic33]

All the people I know called Peter Sutcliff have been perfectly well accepted in society. I even play bridge with one.

I was talking about in the days that followed the arrest. Failed to make that clear.
But who was guilty for the actions carried out against these two people.

 

[MrHappyCyclist]

Do you mean have a dedicated YouTube channel managed by someone

That would be one way to do it, but it would be limited to YouTube videos then.

where videos which are deemed worthy by a group of reviewers are then favourited on that channel?

I was thinking more of self selection, but I suppose a panel of reviewers might help to improve quality. There would need to be a clear set of criteria for selection, and also some criteria for membership of the review panel.

 

[MrHappyCyclist]

Whilst repeat offenders are fairly small, I've little doubt that that will go up in future with the ongoing explosion in the number of cyclists using cameras (and other road users for that matter).

That could be true; especially in London. Myself, I typically see four or five other cyclists in total on my 25 mile commute, so it's unlikely. I do have my program, so I could easily check from time to time.

I wonder whether the seriousness of videos is something worth looking at? Maybe more along the lines of what an editorial group rates videos as, rather than just the video owner, and that a set of tags gets generated as a result.

I think the editorial group thing is the best way to go at the moment, as you say, and possibly some tags could come out of that.

Tagging itself is a means of classification, and any classification scheme is meaningless unless you know what it is for. So the tagging scheme, and the semantics (meaning) of the tags, that would come out of an editorial group would be quite specific to that editorial group's purpose. This was one of the reasons why I put hashes on the suggested tags earlier. For example, the word "dangerous" has different meanings depending on the context: in general is can be taken to mean bringing increased risk to life and limb; in law, when used in the term "dangerous driving", it means that there was deliberate intent on the part of the driver; in our panel's context, we might wish it to mean that there is little doubt over the seriousness of the incident as shown in the video. The hash (or other marker) would be taken to mean "this is a tag whose meaning is defined according to a particular scheme".

It looks like origamist has already implemented a classification scheme that could be a good starting point, with "grade A idiot" being the most serious category.

 

[dawesome]

Sunglasses camera catches stupid drivers, copper implies it's the cyclist's fault:


http://www.thisishullandeastriding....ivers-sights/story-14322545-detail/story.html


Chief Inspector David Rawding, from Humberside Police, said: "We are aware that many cyclists are fitting cameras on to their cycles, however at this time it is unclear how much this will support police investigations.
"I would advise cyclists looking at investing on their bikes and riding apparatus to invest in reflective clothing, cycle helmets, high visibility lights and decent D-locks to ensure their safety and the security of their property."

 

[growingvegetables]

Do you mean have a dedicated YouTube channel managed by someone, where videos which are deemed worthy by a group of reviewers are then favourited on that channel?

That would be one way to do it, but it would be limited to YouTube videos then.

A blog to which the panel of reviewers all had editing access? That at least could have videos from Vimeo and other sources? And the tagging?

Aye - and maybe focus on one message? +1 for "If only we had Roadsafe up here."

I think the editorial group thing is the best way to go at the moment, as you say, and possibly some tags could come out of that.

+1

"Dangerous" - learning from the fair few videos now under my belt, I think I've now got a fairish "feel" for how
- the guys in West Yorkshire Police measure the risk of an incident - above a certain level, and they write and deliver letter by hand, and take the opportunity for a "wee conversation" about safe driving;
- the management at First Leeds assess a bus driving incident - above a certain level, the driver is taken off the road for extra training.

Just throwing that in as possibly one way of helping define "#dangerous"?

 

[dawesome]

Does anyone invest ON bikes? rather strange comments, anyway.

 

[Riding in Circles]

Chief Inspector David Rawding, from Humberside Police, said:
"I would advise cyclists looking at investing on their bikes and riding apparatus to invest in decent D-locks to ensure their safety and the security of their property."

Is he saying it is ok to beat dangerous drivers with a D-Lock??

 

[gaz]

Is he saying it is ok to beat dangerous drivers with a D-Lock??

That is how i see it.