Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Identify car crash editorial anti-patterns using NLP

visionzeroreporting.com

Post reply on HN
Print view View original

151–160 of 213 posts

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#151

Post by chiefofgxbxl »

Tarrosion wrote: post #118

A few months ago I wrote a short Twitter rant about a Boston Globe article that described a pedestrian being struck by a driver really poorly. Thread: https://twitter.com/evanjfields/status/1387131251811831812/ Article: https://www.bostonglobe.com/2021/04/27/metro/woman-28-seriou... To my surprise, this tool finds no problems with the article and gives it a B. (I'm really tickled by the tool, like the idea, but based…

Thanks for linking this issue, and I'll certainly look into it. I agree with your analysis on Twitter: manually annotating the article should yield the following issues:

1. Focus: she was struck by a Toyota Corolla (emphasis is placed on the VRU)

2. Object: she was struck by a Toyota Corolla (agent is referred to as an object, rather than a person, e.g. driver of the Corolla)

3. Object: a 2010 Toyota Corolla being driven by a 29-year-old Brookline man, was driving east (again, the wording personifies the vehicle instead of assigning agency to the driver heading eastward)

I will need to debug this particular example further, but it appears the "2010 Toyota Corolla" and "Toyota Corolla" are not being classified as CARLIKE, a label I trained an NER model on to help deal with all the ways you can name a vehicle: year-make-model, make-model, year-model, model, model-ish, generic terms like truck/pickup/pick-up etc., short-hands like Chevy instead of Chevrolet, etc.

Furthermore, the tool is not identifying "woman" as a VRU. It's a little bit ambiguous because while "[she] entered the roadway", it's not 100% clear that means she is a VRU: a driver/vehicle can enter a roadway too.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#152

Post by markstos »

karaterobot wrote: post #69
markstos wrote: post #55

Earlier quoted context omitted.

I've been employed as a newspaper editor and am familiar with libel laws and the ethics of journalism. What's been advocated for is accurate reporting, not conclusion-jumping. Currently there are headlines like this: "Car hits cyclist" Why is a victim a person but the actor is an inanimate object? They don't say "car hits bike". A more accurate headline would be "Car driver hits cyclist". The other way to improve rep…

> Why is a victim a person but the actor is an inanimate object? To humanize the victim without ascribing blame, when blame is unknown. For the same reason you wouldn't have a headline that read "Alice killed Bob with her car": even though that may technically be true, it's needlessly critical. Needless from the perspective of an editorially neutral party, like a newspaper, but not from the perspective of an activist…

Consider when a terrorist drove a truck into a Christmas market:

https://www.google.com/search?hl=en&q=Christmas%20market%20k...

Most headlines were like "Lorry kills 12 at Christmas Market", like there was a killer truck on the loose.

Here's an alternate framing of an article where a car is a fatal weapon:

"Driver plows car into pedestrians at German bus stop, killing one" https://www.dw.com/en/driver-plows-car-into-pedestrians-at-g... The lead clarifies that "no assumptions should be made at this point".

It's factual that a man drove a truck into a Christmas market and another person drove into a pedestrian at a bus stop. That can be reported without an assumption of guilt without re-framing the story as if the cars drove themselves.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#153

Post by sokoloff »

dionidium wrote: post #59
bluGill wrote: post #26

Earlier quoted context omitted.

One way to make roads safer is to make better drivers. There is a lot of training and continuing education that can be done to make better drivers. It is far too easy to drive, and once you get your license it is very hard to lose it even as your body degrades to the point you cannot see. People think of bad driving tickets (mostly speeding) as just a price you pay to drive. I've never heard of someone getting a tail…

> People think of bad driving tickets (mostly speeding) as just a price you pay to drive. Furthermore, we literally use the phrase, "never got so much as a speeding ticket" to suggest that speeding is the epitome of a minor, inconsequential infraction, when the reality is that we know speeding is a significant cause of and contributor to death, destruction, and injury in automobile crashes.

Every speeding ticket that I’ve gotten in the US was issued on a controlled access road with a speed limit of at least 55mph. None were for as much as 20 over. Most were from pack travel where deciding to travel at the posted limit would result in far more passing and aggregate risk to roadway users.

I do believe all of those are minor infractions (and of course they were inconsequential, except for the delay and expense [if convicted] of the ticket). Revenue enforcement here is mostly done on controlled access highways; it’s shooting fish in a barrel.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#154

Post by chiefofgxbxl »

MauranKilom wrote: post #119
chiefofgxbxl wrote: post #62

Earlier quoted context omitted.

Respectfully, I think you're demonstrating the sort of thinking the corpus of research surrounding this seeks to shine a spotlight on: blaming the victim, rather than focusing on safe and forgiving design. Should our built environment be such that falling down in public should end your life? Do people deserve to die simply because they want to cross the street? Why do we design residential neighborhoods where parents…

Calling them "editorial anti-patterns" because they don't frame the discussion how you would like it to be framed is... well, an antipattern in itself. Let me be clear: Your goal of improving road safety is laudable, and the aspects your tool highlights are conceivably a way to achieve that. But priming visitors to expect "editorial anti-patterns" and then instead presenting "insufficiently biased-to-the-pedestrian"…

Thanks for your feedback!

I certainly have a lot more work to do over the next few weeks to improve the tool, and I think much of your comment will be addressed once I place a lot more weight on Framing (see my other comment: https://news.ycombinator.com/item?id=28842314).

For your "driver hit the cyclist" example, I would like to avoid confusion that there was a physical altercation like a fist fight between the driver and cyclist. However, I stand by the sentence structure, because I don't want to personify or give agency to vehicles (e.g. "vehicle hit the cyclist") when it's the driver who is in control, and focusing on driver calls into question more thematic elements like distracted driving, speeding, visibility and lighting conditions, etc.

Is there a way you'd recommend rephrasing "driver hit the cyclist" to satisfy both our suggestions?

> Calling them "editorial anti-patterns" because they don't frame the discussion how you would like it to be framed is

To be clear, I'm not trying to force my desired framing on authors or inject bias into articles. I am simply working off of the research I've seen, such as Editorial Patterns in Bicyclist and Pedestrian Crash Reporting [1] and all the other citations in that paper, that there is a very real and measurable effect on the language used in these articles and the readers' perception on blame and preventative measures.

However, I am certainly thinking about the feedback I've received from everyone in this thread.

> If you want to improve road safety, you could also recommend report on the maintenance status of the vehicle (e.g. were the brakes well-serviced?).

That sounds like an interesting idea. I know that vehicle status can contribute to crashes, and legislation has been enacted to improve this. For example, all vehicles now require tire-pressure monitoring system (TPMS) because improper tire pressure was leading to lots of crashes. Back-up cameras are required because people, particularly small children, were being run over when in reverse. States have different requirements for periodic vehicle safety inspections (in New York State where I live it's once per year).

If a crash is being reported involving older vehicles, maybe journalists could mention if those vehicle predate certain safety mandates. For example, vehicles before 2007 in the U.S. may not have TPMS [2]

[1] https://www.researchgate.net/publication/330975590_Editorial...

[2] https://en.wikipedia.org/wiki/Tire-pressure_monitoring_syste...

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#155

Post by Tarrosion »

chiefofgxbxl wrote: post #151
Tarrosion wrote: post #118

A few months ago I wrote a short Twitter rant about a Boston Globe article that described a pedestrian being struck by a driver really poorly. Thread: https://twitter.com/evanjfields/status/1387131251811831812/ Article: https://www.bostonglobe.com/2021/04/27/metro/woman-28-seriou... To my surprise, this tool finds no problems with the article and gives it a B. (I'm really tickled by the tool, like the idea, but based…

Thanks for linking this issue, and I'll certainly look into it. I agree with your analysis on Twitter: manually annotating the article should yield the following issues: 1. Focus: she was struck by a Toyota Corolla (emphasis is placed on the VRU) 2. Object: she was struck by a Toyota Corolla (agent is referred to as an object, rather than a person, e.g. driver of the Corolla) 3. Object: a 2010 Toyota Corolla being dr…

Interesting context, thanks! Cool tool, will be glad to see it get tuned further :)

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#156

Post by avianlyric »

smoldesu wrote: post #147
burnished wrote: post #136

Earlier quoted context omitted.

Its more the realization that you aren't safe, and that your safety is not only not a priority but somehow actively repugnant, all due to you being on a bike. If you are a kid driving isn't an option. If you don't feel safe on the roads as a pedestrian or bicyclist then your options are rather limited. This was my own experience, and if I had been less reckless I would have gotten less exercise. As it was, telling so…

Sounds like you should cycle in safer places then? Airline pilots don't complain because highways aren't a safe place for planes to be, and similarly it can be pretty dangerous to ride a bike in a place where cycling is an afterthought. Shifting the blame back to the drivers doesn't change that, if anything it tarnishes the reputation of cyclists even further. I get a similar feeling when I see people on Hacker News…

> Bad infrastructure is an infrastructure issue. Trying to gussy it up as a social one is going to be a really tough case to…

Last I checked infrastructure spending was mostly dictated by social policies. Either that or we’ve had generations of truly incompetent infrastructure planners who have all failed to realise that drivers in cars aren’t only people who need to get from A to B.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#157

Post by smoldesu »

avianlyric wrote: post #156
smoldesu wrote: post #147

Earlier quoted context omitted.

Sounds like you should cycle in safer places then? Airline pilots don't complain because highways aren't a safe place for planes to be, and similarly it can be pretty dangerous to ride a bike in a place where cycling is an afterthought. Shifting the blame back to the drivers doesn't change that, if anything it tarnishes the reputation of cyclists even further. I get a similar feeling when I see people on Hacker News…

> Bad infrastructure is an infrastructure issue. Trying to gussy it up as a social one is going to be a really tough case to… Last I checked infrastructure spending was mostly dictated by social policies. Either that or we’ve had generations of truly incompetent infrastructure planners who have all failed to realise that drivers in cars aren’t only people who need to get from A to B.

Sounds like you've answered your own question there.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#158

Post by avianlyric »

dymk wrote: post #9

This is neat, but something bugs me about the framing of the intended goal versus how you propose to get there. If the intent is to raise public awareness and to put pressure on leadership to make roads generally safer, why do almost all the corrections follow a pattern of shifting blame from the individual pedestrian onto the individual driver? Many of the suggested fixes are just their own form of counterfactual -…

In all of these scenarios the driver is the better protected and less vulnerable participant.

Ideally we should be improving infrastructure to ensure that the risks created by high speed lumps of metal is reduced as far as practically possible. But until that happens, we have to acknowledge that in the majority incidents drivers are only ones capable of changing the outcome. Additionally drivers should take one a greater responsibility when operating vehicles in areas where they might encounter more vulnerable road users. Vulnerable road users will always end up paying a substantially greater price in any incident, so it’s reasonable to ask drivers to be more careful to avoid those incidents.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#159

Post by Zak »

rtlfe wrote: post #142
Zak wrote: post #89

Earlier quoted context omitted.

What I remember reading about the impact of speed limits is that reducing them only affects the slowest quartile or so of drivers while changes to road design such as reducing road width and visibility slow down almost all drivers. Reducing the speed of some drivers and not others increases aggressive driving and probably isn't good for safety. I don't take the position that reducing speed is a good answer for every…

> I don't take the position that reducing speed is a good answer for every high-risk road situation Why not? There's clear evidence showing that crashes at higher speeds are much more deadly. IMO any street that allows pedestrians and bicycles should be designed for 20mph car traffic. If you want to drive faster than that, do it on a limited-access highway.

Limiting access to the road, e.g. using pedestrian overpasses, fencing, segregated sidewalks and bike lanes, etc... is among the options I had in mind. I don't think one solution fits every problematic road.

Re: Show HN: Identify car crash editorial anti-patterns using NLP

#160

Post by Karrot_Kream »

sokoloff wrote: post #153
dionidium wrote: post #59

Earlier quoted context omitted.

> People think of bad driving tickets (mostly speeding) as just a price you pay to drive. Furthermore, we literally use the phrase, "never got so much as a speeding ticket" to suggest that speeding is the epitome of a minor, inconsequential infraction, when the reality is that we know speeding is a significant cause of and contributor to death, destruction, and injury in automobile crashes.

Every speeding ticket that I’ve gotten in the US was issued on a controlled access road with a speed limit of at least 55mph. None were for as much as 20 over. Most were from pack travel where deciding to travel at the posted limit would result in far more passing and aggregate risk to roadway users. I do believe all of those are minor infractions (and of course they were inconsequential, except for the delay and exp…

That's another part of this issue. Traffic enforcement in the US is more ideological than it is utilitarian. Automated enforcement tools like red light cameras and speeding sensors run into intense political pushback in the US. Speeding tickets are usually ways to boost revenue instead of actually trying to deter speeding behavior. Traffic stops are arbitrary and usually based on the political and ideological goals of the Police department and current local government party instead of actually trying to reduce traffic incidents. In fact, as far as I know, traffic incidents are mostly used to drive infrastructure changes instead of really feeding into changing traffic policing.

There's real political pushback from trying to actually decrease traffic incidents because actions to decrease these incidents result in higher overhead for current drivers, something historically lightly enforced.

Post reply on HN