Session 9: Local SPF Use- Iowa Michael Pawlovich, Ph.D., P.E. Iowa Dept. of Trans., Traffic and Safety
Iowa “SPF” Development Don’t really use “SPF”s, per se, currently – do use CRFs/AMFs Have developed models, mostly for evaluation of past countermeasures Don’t use EB… – went from classical FB – essentially same difference but Iowa resources allowed – development continues July 29 and 30, 20092
Iowa “SPF” Examples Examples of Iowa FB use: – Bayesian Intersection Ranking (past) limited # of sites demonstration case – 4 3-lane Before/After Evaluation Study (past and published) could use as an SPF but haven’t had need – 2 3-lane Before/After Evaluation Study (ongoing) July 29 and 30, 20093
Iowa “SPF” Examples More examples of Iowa FB use: – Impacts of Bypasses Before/After Evaluation Study (ongoing) – Bayesian SICL (network screening) – intersection and segment (1 st ongoing, 2 nd future) – Comparables/Expected Values (future) – Alternate considerations for location – not site-based but crash-based (future) July 29 and 30, 20094
Iowa SPF Thoughts Methodologically, we can develop prediction models or SPFs – begin from engr. problem and site data – develop stat. models faithful to engr. concepts and include parameters interpretable in the engineering context – Model parameters estimated using fully Bayesian (hierarchical) methods July 29 and 30, 20095
Iowa SPF Thoughts We are trying to develop models/SPFs useable over a reasonably wide range of site types – base estimates on datasets with a diversity of site attributes – model over these diverse attributes July 29 and 30, 20096
Iowa SPF Thoughts For example, if we wish to investigate one type of site vs. another we can use the same SPF as long as the data and model properly contain the means Limitation not software or stat. tools but rather that needed data not available to permit estimating canonical SPFs – working towards July 29 and 30, 20097
Conclusion Data improvement crucial – canonical SPFs – specific SPFs – analyses in general Learning and improving as we go – "If we knew what it was we were doing, it would not be called research, would it?“ - Albert Einstein July 29 and 30, 20098
Thank you July 29 and 30, 20099