Difference between revisions of "Staff Priorities - Masha"

Revision as of 17:31, 26 August 2015

Responded to Lukas H. at ETH regarding CSEP testing center
Investigating disk usage on csep-op
Trac ticket #360: Install GEAR model by Peter Bird for high-resolution Global testing region
- Generated forecast by acceptance test, takes ~3 days
- Contacted Peter asking to shorten the run for the purpose of acceptance test
- Compiled model with '-parallel' option, getting failure, contacted Peter
Max's data request for the CSEP workshop:
- Posted existing csep-op results
- Posted Canterbury 1-day and cross-group T and W test results
- Set up 3 custom experiments per Max request:
  - T and W tests for 3-month vs Md2 3-month cross-group evaluation
  - Error evaluating 30-min vs 1-day forecasts, and 1-day vs Md2 1-day groups, need code changes

Max's data request for the CSEP workshop:
- Code changes to allow for 30-min vs 1-day and 1-day vs. 1-day cross group evaluation with T and W tests
- Set up custom experiment for best performing models in RELM aftershock vs. 3-month vs. 1-day models for CA
#358: S and M evaluation tests should not be invoked when there are no events observed for the test

Blocked: No

Followups:

With John Y.:
- csep-op: Added my email to recipients list for Oceanic Transform Faults status email, got email sent to my account, but not to the mailing list.
- CruiseControl installation for csep-cert

Trac ticket #353: Optimize RELM N-test by avoiding event to bin look up when masking bit is set for the whole forecast
Trac ticket #346: Evaluate Oceanic Transform Faults forecasts since 2012/01/01

@@ Line 18: / Line 18: @@
 === Today ===
 * Max's data request for the CSEP workshop:
+** Code changes to allow for 30-min vs 1-day and 1-day vs. 1-day cross group evaluation with T and W tests
+** Set up custom experiment for best performing models in RELM aftershock vs. 3-month vs. 1-day models for CA
 * #358: S and M evaluation tests should not be invoked when there are no events observed for the test