search

sunlightpolicy / State-Open-Data-Census

Working towards a US State Open Data Census
11 stars 3 forks source link

Define Dataset: Reportable Motor Vehicle Crash Data (Transport & Infrastructure) #27

Open emily878 opened 9 years ago

emily878 commented 9 years ago

Thank you so much for this suggestion, @dsmorgan77! Could you help us define the essential substantive elements of a "Reportable Motor Vehicle Crash Details (fatal and non-fatal) and related enforcement" dataset? What are the components that it must minimally include? Is one of the NY datasets more inclusive or useful than the others if we wanted to hold it up as a model?

https://data.ny.gov/browse?q=%22motor%20vehicle%20crashes%22&sortBy=relevance&utf8=%E2%9C%93

dsmorgan77 commented 9 years ago

Any good crash data set includes at least the elements of the model minimum uniform crash criteria (http://mmucc.us). That means information about the crash, the vehicle(s) and person(s) involved. NY chooses to release those aspects as individual (but linkable) files. All are valuable. On Mar 6, 2015 4:31 PM, "Emily Shaw" notifications@github.com wrote:

Thank you so much for this suggestion, @dsmorgan77 https://github.com/dsmorgan77! Could you help us define the essential substantive elements of a "Reportable Motor Vehicle Crash Details (fatal and non-fatal) and related enforcement" dataset? What are the components that it must minimally include? Is one of the NY datasets more inclusive or useful than the others if we wanted to hold it up as a model?

https://data.ny.gov/browse?q=%22motor%20vehicle%20crashes%22&sortBy=relevance&utf8=%E2%9C%93

— Reply to this email directly or view it on GitHub https://github.com/sunlightpolicy/State-Open-Data-Census/issues/27.

waldoj commented 9 years ago

It's so delightful that there's already a standard. :) Such a rarity. And it's got "minimum" right in the name!

waldoj commented 9 years ago

I'm not how I should rate those states that only provide statistics, as opposed to incident-level data. Do we not count it at all? Or do we just ding them on whether it's complete? Given the huge gap between statistics and data, I worry that just lowering their score for completeness is insufficient to represent the scale of the problem here.