...

Soujanya Jarabana and Mike Satteson

Questions from Law Enforcement Agencies:

• CRU business unit provided email and remote support to law enforcement agencies to resolve Validation errors in the DT4000 crash form. 

• Received 15 LEA questions in the last 4 weeks involving 1 Hit and Run Fatal Crash, 2 crashes about Crash diagram, 3 Questions about Non-Reportable Crashes.

...

 Transmission failures:

• We addressed 19 transmission failures between March 24th until now after providing solution to the agencies.

  • 5 are transmitted

  • 12 are transmitted as other Document (Duplicate Document Number)

  • 2 are pending with Agencies.

  • A Persistent source of Duplicate Document failures

    Jira Legacy
    server System JIRA serverId 27e84956-8aca-34a6-b686-31eab275489e key CRUBIZ-317

    appears to have been fixed after both the CRU and Badger TraCS worked with the agency. Between 15 January and 12 April, we saw 19 of these document number reuse failures, 9 required CRU follow-up to be corrected.

DT4000 Fatal Crashes:

Quality Check process for DT4000 fatal crashes is performed which comes in Resolve. The process starts with creating a Jira issue for DT4000 basic check and then and assigned to the FARS team,

...

Currently, we are missing 1 crash from Washington County SO 

...

 Enhancements to Crash Database & Resolve System

• Currently in ready for production for planned deployment on 2023-05-03 which includes.

...

Ken Matthews

Reference Point Coding

...

Asadur Rahman

Jacob and Erickson coded since 03/31:

Jacob: 2,359

Erickson: 1,629

Total: 3,988

RP coding needed for 2022: 43,784

Crash Data Analysis and Reporting

Valerie Payne and Rebecca Szymkowski

• Wisconsin Traffic Crash Facts

Scheduled Topics and Briefing documents

2021 Crash Data Refinalized to Add RPs

2021 RP Coding was completed in early April. The TOPS Lab worked to create a new version of the 2021 finalized data that differs from the initial version only in the added RP fields. The CRU checked the results to verify that this was the case.

...

This process had to go through a couple of iterations because of challenges posed by DMV customer number assignment changes (which do not increment the crash record number). The original 2021 finalized data had 20,271 crashes with RPs. The post freeze finalization added 33,768 RP for a total of 54,039 RP in the refinalized data:

julia> freqtable(x, :old_has_rp, :new_has_rp)
2×2 Named Matrix{Int64}
old_has_rp ╲ new_has_rp │ false   true
────────────────────────┼─────────────
false                   │ 74257  33768
true                    │     0  20271

The check process verified successful refinalization of both flat file and SAS library extracts. According to a cell by cell comparison, the updated version of these data is identical to the initial version in all respects except for

• Data Record Number, Record Timestamp, and the Agent responsible for the record change (RECDNMBR and RECDDTTM in all tables and OPCODE in crash)

• RP data in the roadway. table (additions only, no changes were made to RP data that had been entered when the data were finalized).  These fields are:

  • RPRTEID

  • RPRTETY

  • RPRTENB

  • RPRTEDIR

  • RPTYPE

  • RPNMBR

  • RPDIS

  • RPLINK

  • RPLKOT

  • ALTLAT

  • ALTLON

  • ALTSRC

• The order of records in the two versions may be different, but records can be matched exactly by appropriate sorts on a single key field in each table. (Order was preserved in the extract to dataframe datasets that I had examined earlier.)

We made the SAS library version of the these refinalized data this week as test libraries. Other at large version of the 2021 finalized crash data re unchanged pending workgroup decision

• Does the workgroup approve the release of the 2021 refinalized data as our 2021 crash statistical data set?

2022 Finalization

We turned our attention to 2022 Crash data Finalization, which we need to complete by the end of May. We were not able to keep up the continuous data cleaning in response to daily batch edit checks that we had begun to realize in late 2021. In part this is because Tejal had done much of this work, but we are also working to rebuild our data quality and data query infrastructure to use Jira. We should still be able to get the data clean for critical checks, but our quality measures against important edit checks will likely be poorer than what we had been able to achieve for 2021.

Changes to Edit Checks

We discontinued 12 FARS related edit checks. FARS checks are not done as part of our in development new case management system, which is being implement in Jira. We’ve begun to implement checks in that system. Currently we have 31 checks implemented in that system, and we typically clear those each day.

Moved Narrative: Bicycle Mentioned, but not a Unit from Critical to Important. This had been an Informational check prior to 2022-01-13.

Current Status

For checks run on 2023-04-27 against 2023-04-26 production 2022 data. We had

The tally for critical checks is

Based on past experience, most of these critical check hits can be cleared without laborious agency follow-up. We ought to be able to clear them in time to finalize with data that is clean of critical checks.

Close

Informal Closing Discussion