Data Quality Today - PowerPoint PPT Presentation

1 / 9
About This Presentation
Title:

Data Quality Today

Description:

Comments from the Offline Shifters have lead to slight changes in the procedures, ... on data quality process rather than on substance for my tastes. On one hand, procedure ... – PowerPoint PPT presentation

Number of Views:15
Avg rating:3.0/5.0
Slides: 10
Provided by: mogu
Category:
Tags: data | lead | quality | today

less

Transcript and Presenter's Notes

Title: Data Quality Today


1
Data Quality Today
ADM Meeting July 25th, 2003
  • RecoCert Status
  • RecoCert Offline Shifters
  • RecoCert Results
  • Punch List Changes

2
RecoCert Status
  • The Offline Monitoring Process has been reduced
    to a few of steps. Example
  • setup D0RunII p14.03.01 d0cvs sam d0tools
    d
  • Setenv RECOCERT_WORKAREA somedir for the results
  • RecoCert workarea home is available on clued0
    this is what was holding us up on June 27th.
  • recentDataInSam
  • runRecoCert --run178148 --versionp14.03.00
    --num1000
  • verifyRecoCert --run178148 --versionp14.03.00
  • Makes plots and OverLay Plots (the overlays
    dont work yet) and comparison plots are
    available on the DQ web page.

3
Offline Shifters RecoCert
  • Expansion of the SAM-Shifter role to include
    Offline Data-Quality Monitoring providing
  • backup for the online gear of the Data-quality
    protection
  • Mechanism for catching problems that occur in
    Reco
  • General Plan for RecoCert
  • Shifters would run it and look at recent output
    from Reconstruction Farm.
  • Record results in the Online Logbook and possibly
    Run Qual DB.
  • Note that the monitoring occurred and any
    discrepancies that were uncovered.
  • Discrepancies result in email to Diehl for now
    who (in normal hours) will contact the
    appropriate expert.
  • RQC maintains a Punch List of problems and
    documents expert follow-up.

4
Offline Shifters RecoCert
complete
  • Goals Established in Workshop run it at least 5
    times by July 6th and at least 15 times in July.
  • By July 6th we had completed the RecoCert process
    once.
  • In Run 178148 (done) there was a little blip in a
    tracking plot but otherwise it seemed quite good.
  • As of July 17 1400 we have three completed
    RecoCert processes and some in-progress by
    Offline Shifters R. Illingworth and
    S. Crepe.
  • Runs 178737 (done), 178765 (done), 788, 789,
    796.
  • As of July 24 1600 we have 11 completed RecoCert
    processes and many in-progress by Offline
    Shifters
  • Comments from the Offline Shifters have lead to
    slight changes in the procedures, particularly
    with regard to the run selection process.

5
RecoCert Results Run 178737
  • Most histograms looked good or typical.
  • Tracking plot has some evidence of excess z0
    vertices. I was under the impression that this
    sort of thing should be somewhat rare now.
  • Status punch list for verification of the nature
    and cause of the problem.

6
RecoCert Results Run 178737 65
  • Most histograms looked good or typical.
  • EMF plot used to have much more in region EMFgt0.8
    as if the electrons were included in the jets.
    Now it looks like they are disincluded.
  • This may not be a problem but instead reflect a
    change in definition.
  • Status Punch list (until we understand that the
    change is intentional).

7
The Jet Thing
  • Run 178721 vs 178722
  • 178721 is trigger version 11.04
  • 178722 is version 12.10
  • Suggests the change in EMF occurred with change
    in trigger.
  • But the number of jets per event is the same in
    these two runs.
  • We didnt punch list the change in shape of ET
    spectrum (Good jets are a close match with
    eachother).
  • Might be the non-linearity corrections which are
    now applied in L3.
  • Might be the tighter EM selection in EM trigs.

8
Punch List
  • Can be found at
  • http//www-d0.fnal.gov/computing/data_quality/punc
    h/punch_list.txt
  • Three items
  • Item 12 Date July 17th. The EMF thing in
    recent runs.
  • Item11 Date July 17th. The z0 vertex thing in
    Run 178737.
  • Item 4 Date April 11th. Readout Errors in SMT
    from end of Jan. (2003) shutdown to Mar. 10th
    effected the reconstruction. Runs sould be graded
    as "Not perfect but analyzable". Can this be seen
    in RecoCert plots? Call List HTD. Action Ron
    Lipton is still worried about this July 17, 2003.
    Resolved Date and Comment

9
Summary
  • RecoCert Status
  • It goes. It makes plots. The plots are documented
    but the documentation could be even better.
  • RecoCert Offline Shifters
  • They are doing it!
  • RecoCert Results
  • Three successful completions. More underway.
  • Punch List Changes
  • Two items added to the Punch List making the
    total 3 active items.
  • Next week Run Quality Database news.
Write a Comment
User Comments (0)
About PowerShow.com