To do 28/03

  • jet truth 4vectors into ntuple dump
  • truth lepton 4vectors into ntuple dump
  • Use selected jets for LAr hole jet veto ----> DONE
  • EM not EMJES for overlap removal jets

To do 13/03

Zcc -> ntuples without the HFOR cut


> This is to see if the charm template is different to the one we currently use.

dilepton vs OR of single/dilepton

Run on data and check for both electron and muon channels is there no difference after pt and eta cuts? Check this for each run period.


> This is to see whether we care at all about this trigger question. If it gives the same number of events after pt and eta cuts then we can use the single lepton scale factors (dilepton don't exist).

Event weight -> Check that it's the same in the dumper as in the analysis after applying the efficiency corrections.


> This looks buggy! The disagreement between data and MC should be improved by applying the weight.

  • Fix MET DONE
  • Fix Pileup reweighting
  • Apply btag eff sf??
  • PAULS BINNING

To do 12/03

  • Fit tagger variables
  • Fix Z window bug
  • Verify MET DONE
  • Binning DONEISH
  • (compare periods, pileup etc)

To do 07/03

  • Change/sort out triggers ---- DONE
  • Systematics!
  • Further tidy of code
  • fixed control plots ---- DONE

To do 02/03

  • Cutflow comparison MC11c
  • Tidy up zbjetskel/zbjetbase
  • Systematics
  • Control plots for tuesday

To do 09/02

  • Add MET branches & histos (were there before from Higgs)
  • Test/run 11c
  • Check MET cuts, isBad, reject event flag etc
  • Add MV1 branch to ntuple

To do 03/02

  • * tidy up code, remove unused code and histograms *

  • email matt

  • cut flow comp, side by side for Z, Z + Jets, Z + bjets, for all groups

To do 20/01

  • ADD event weight
  • Check jet cleaning branches, check dumped selected jets
  • Validate ntuple info, run over 10 000 ZeebbNp0 events
  • Z dump, before Z mass window: after dimuon/dielectron opposite charge
  • Remove extra mc truth info from dumper
  • Think about implementing systematic shifts in dumping code

Priority

  • Fix Jet problem to allow comparison of cutflows
  • Estimate total ntuple outputs for Z+b and Z+jets with reduced #histograms ------------> CUT OUTPUT SIZE IN HALF SO FAR
  • Tidy up code, increase efficiency
  • Produce proper plots of tagging distributions (binning issue... re-run on grid once jet problem is solved)

To do list 16/01/11

  • Ping Matt about the yields, they look problematic. ----------> DONE, HE IS HAVING ISSUES TOO, YET TO LOOK

  • Cut flows, compare them to Lecce, if they look remotely reasonable (order of magnitude) then it's worth comparing them -------------> E/MU FINE, PROBLEM WITH JETS

  • Histogram clean-up: bin everything you've never looked at (copy to a backup file so that you can always put it back in) ---------> PRIORITY

  • Histogram plotting: Be in charge of that code, will be easier with far fewer histograms. -----------> PRIORITY

To do list: 12/01/11

  • fix plotting script and produce tagging distributions, extra plots are fine (ignored) but give me the names of the ones I need. ---------> STILL NEED TO FIX FOR JFITTER PLOTS

  • Run for latest cutflows and plots (needed by tomorrow when everyone else will post). -------------> DONE, JETS NEED FIXING

  • Run ntuple dumping job. ----------------> DONE

  • Create a twiki (you should have a personal one at CERN that you could use) - plots and to-do lists can be maintained on there, then we can refer to that when discussing the next steps, record what's been done, etc. -------------__> DOING ; )

  • Use a logbook, especially when changing the code. -------> NOTED ; )

  • Remove unneeded histograms (affects performance and merging). ---------> PRIORITY

  • General analysis code cleanup to minimise the amount of code ran on the grid and its complexity (maintenance and turnaround affected a lot by this, performance might improve as well depending what you find). -------> PRIORITY

  • Cleanup of the plotting script, on a longer time scale you should rewrite this from scratch, stealing snippets from the existing one as you need them. -------> LOW PRIORITY

To do list: 23/12/11

  • produce the inputs for fit studies --------> DONE
  • long-term data processing model (ntuple-dumping), and trimming the job down in terms of #histograms ------> MORE TRIMMING, CLEANING HISTOS NEEDED
  • the actual analysis, systematics ...

  • Jobs:

  • Run over data and MC and get all of the histograms of fit inputs (all JFCNN, SV0 mass, differentially). I don't want an SV0 tagger cut if I'm looking at JFCNN weights. --------> DONE

  • Can you dump TTrees of those fit input variables easily (as well as histograms)? -----> DONE

  • ntuple dumping:
  • trim away stuff that's obviously not needed, if you're unsure then keep it.

  • Try to dump the ntuples after the full selection (they'll be the smallest number of events) - if you've implemented it after the Z selection that's also fine and the long-term plan.

  • try dumping for ~0.5fb-1 of data just to make sure it's not really massive!

  • Re-include histograms - if it's the number of histograms that's the problem, remove those histograms that you don't need (copy to backup file). For the differential plots, be selective about what you histogram differentially, it's useful to have some control plots as well as the fit inputs, but not absolutely everything!

  • QCD systematic - implement!

  • Lowest priority is a full update to the Higgs analysis status, this will be much easier and more efficient when you're back at CERN, or maybe already in Glasgow?

-- AaronBundock - 17-Jan-2012

Edit | Attach | Watch | Print version | History: r14 < r13 < r12 < r11 < r10 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r14 - 2012-03-28 - AaronBundock
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Sandbox All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback