5 ms·

Hey HN, I'm the lead person who did the original data collection. We were previously featured when I made the initial site here: https://news.ycombinator.com/it

Hey HN, I'm the lead person who did the original data collection. We were previously featured when I made the initial site here: https://news.ycombinator.com/item?id=16840237 https://news.ycombinator.com/item?id=16840237 Happy to answer more questions people might have about methodology or the analysis in the original study.

olau 7y ago
I know this is perhaps besides the point: but I'm surprised that traffic incident reporting is so low - but then aren't you scraping more intellectual news media than what I would imagine most people would be reading? Not very familiar with the English/American news situation.
- Arnt 7y ago
  You can do a low-quality check easily. Read some medium that you suspect will be different from NYT/G for a week and count how many traffic accidents were reported in a particular locale, and how many reports there were for each. Then call the police and ask how many there actually were in that week. I can guess what you'll find: What's chosen for reporting is news or otherwise noteworthy events. Traffic accidents may be a cause of death, but that doesn't make them either news or noteworthy. Traffic jams and their causes are an everyday occurence, even if the cause involves someone being driven off in an ambulance.
- owenshen24 7y ago
  I would have loved a more comprehensive search of media. We opted for NYT/G because they had the friendliest APIs given our time constraints.
samat 7y ago
I am very interested to see - comparison of the share of deaths to the share of google searches; - comparison of google searches to the media coverage of deaths. Would it be possible for you to generate those two?
- owenshen24 7y ago
  I'm not exactly sure which analyses you'd like to see, but all the data is here in csv format if you want to run your own visualizations: https://github.com/owenshen24/charting-death-analysis/tree/master/data https://github.com/owenshen24/charting-death-analysis/tree/m...
charlescearl 7y ago
Since you are around, the thesis > But is there evidence that such a disconnect exists between what we see in the news and what is reality for most us? is an interesting one, particularly in this case. The leading cause of death for African American males 15 - 35 in 2015 was homicide according to CDC statistics. It seems that’s a very important news worthy focus, especially with respect to policy debates that shape the way US cities are run. Even including caveats in your report would bring an important and nuanced view. I’m curious as to why this was not discussed.
- owenshen24 7y ago
  One thing I wished we had done originally was stratify by age. Commenters on the original post mentioned that it wasn't entirely accurate to average over all deaths, esp. as different things are more likely to kill you at different times in life. The short answer for why we didn't do it is that I didn't realize this until after we finished the project. This was for a university data science project, and we were working on limited time.
  - charlescearl 7y ago
    Especially given it was a class, great work. Appreciate that the GitHub is accessible and transparent.

olau 7y ago

I know this is perhaps besides the point: but I'm surprised that traffic incident reporting is so low - but then aren't you scraping more intellectual news media than what I would imagine most people would be reading? Not very familiar with the English/American news situation.

Arnt 7y ago
You can do a low-quality check easily. Read some medium that you suspect will be different from NYT/G for a week and count how many traffic accidents were reported in a particular locale, and how many reports there were for each. Then call the police and ask how many there actually were in that week. I can guess what you'll find: What's chosen for reporting is news or otherwise noteworthy events. Traffic accidents may be a cause of death, but that doesn't make them either news or noteworthy. Traffic jams and their causes are an everyday occurence, even if the cause involves someone being driven off in an ambulance.
owenshen24 7y ago
I would have loved a more comprehensive search of media. We opted for NYT/G because they had the friendliest APIs given our time constraints.

samat 7y ago

I am very interested to see - comparison of the share of deaths to the share of google searches; - comparison of google searches to the media coverage of deaths. Would it be possible for you to generate those two?

owenshen24 7y ago
I'm not exactly sure which analyses you'd like to see, but all the data is here in csv format if you want to run your own visualizations: https://github.com/owenshen24/charting-death-analysis/tree/master/data https://github.com/owenshen24/charting-death-analysis/tree/m...

charlescearl 7y ago

Since you are around, the thesis > But is there evidence that such a disconnect exists between what we see in the news and what is reality for most us? is an interesting one, particularly in this case. The leading cause of death for African American males 15 - 35 in 2015 was homicide according to CDC statistics. It seems that’s a very important news worthy focus, especially with respect to policy debates that shape the way US cities are run. Even including caveats in your report would bring an important and nuanced view. I’m curious as to why this was not discussed.

owenshen24 7y ago
One thing I wished we had done originally was stratify by age. Commenters on the original post mentioned that it wasn't entirely accurate to average over all deaths, esp. as different things are more likely to kill you at different times in life. The short answer for why we didn't do it is that I didn't realize this until after we finished the project. This was for a university data science project, and we were working on limited time.
- charlescearl 7y ago
  Especially given it was a class, great work. Appreciate that the GitHub is accessible and transparent.