Close Menu
    [OC] US vehicle occupant risk
    Data

    [OC] US vehicle occupant risk


    Hi-res zoomable, color-blind, and absolute-risk versions can be viewed here.

    This is an updated model (with better parameters and imputation) and includes another year of data compared to the previous one.

    Data are normalized by road type so similar roads are comparable.

    Data were compiled using national fatal crash and volume data (FARS 2021-2024, HPMS). Only FHWA functional classes 1-6 are shown (local roads are excluded to make a cleaner visual).

    Posted by jejmcjej

    Share.
    View 30 Comments

    30 Comments

    5. InfraredDiarrhea on

      I’ve done a lot of driving and this map confirms my anecdotal experiences with Memphis, TN and i70 through Missouri. 

    6. wanliu on

      A question on your methodology. Obviously you don’t have deaths on every mile of every road, did you grid deaths and then smooth these to adjacent roads? For example, in Chicago would deaths from accidents on interstate 94 be applied to the surrounding arterial roads?

    7. BrightLuchr on

      The shown image seems quite deceptive. The absolute risk map looks more accurate. At the very least, it needs to be normalized by vehicle-miles traveled. This should clearly show rural roads as being the most dangerous compared with city roads..

    8. mr_ji on

      I-40 between Flagstaff and Albuquerque being so dangerous despite how much less traffic there is than most of these routes really says something.

    9. exneo002 on

      it’s surprising so many metros have high death rates because I heard proximity to hospitals is a big factor in survival (like gun violence)

    11. Fabulous-Grocery2441 on

      For being the state with the most alcohol consumption, Wisconsin is impressively green.

    12. Mid_Atlantic_Lad on

      Memphis does NOT surprise me at all. North Dallas through is interestingly low risk compared to the rest of the city.

    16. pistachiopudding on

      Pittsburgh is a significant outlier of being safe through the whole city without the whole state being safe. Anyone know why?

    17. tlmbot on

      I am perplexed by New Orleans – LA is the craziest place to drive I have ever lived.

      (gotta go to the high res map for it to even really show up)

      huh, I guess it’s just a different kind of accident in NOLA. That place really is different.

      out of town is much more “high velocity bad decisions.” In town is much more…. pot holes/watter filled bottomless pits, drivers stopping and backing up/outright going the wrong way, people randomly in the road… etc. oh, saw a gator on the interstate once. Probably a daily occurrence and I just didn’t pay attention.

    18. M35Dude on

      This is very interesting! Do you have a diff map between years? I’d be really interested to know how much this varies year-to-year.

    19. InvisibleBlueUnicorn on

      in seattle metro area, I-90 divides red-green clearly. North of I-90 is green and south is red.

    21. 15pH on

      Thanks for this interesting and useful data visualization!

      One factor I’m curious about is hospital distance and quality, and how much that matters. It might be interesting to have a hospital overlay and/or data set normalized to hospital distance.

      What are some other factors that you might decouple if you wanted to expand this project?

    22. Dust-Different on

      So the ones who drink the most have the safest rides? What are you guys doing in the south? Meth?

    27. Data_chunky on

      At first I thought it was funny that the Bay Area (Cali) was so green, but then I realized that with all the traffic they are probably going 5mph and it’s not that dangerous. 😂

    28. quickthrowawaye on

      One caution: normalizing by FHWA road class doesn’t necessarily make roads comparable: roads in the same class can differ significantly in design, traffic, and risk. You’ll introduce some variation, probably most notably across state boundaries.