cms data analysis patrick mooney trinity school south bend, in jill ziegler hamilton high west...

CMS Data Analysis

Patrick MooneyTrinity School

South Bend, IN

Jill ZieglerHamilton High West

Hamilton, NJ

Megan GillianLa Lumiere School

LaPorte, IN

Nick SchrockBethany Christian Schools

Goshen, IN

Starting Points

• Work last summer with the tt̅� single muon data set and tutorial by Drs. Sander and Schmidt– dimuon mass– preliminary work viewing plots– preliminary estimation of cuts for top events

• Foundational work with C and C++

Data Updates and Access

• Data set, Monte Carlo background simulations, and Tutorial updated in fall 2014 by Drs. Sander and Schmidt– Added three new background simulation types

• New tutorial data and sample programs uploaded to server by Dr. Dan Karmgard– Access to server through virtual terminal– Individual directory space for each user

• Library created for shared unmodified files by Dr. Karmgard– Holds archive copies of frequently modified files– Holds working library of infrequently modified files– Main program files modified to call data directly from central

library

Histogram Plotting

• Works through subprogram “Plott̅er”• Data or simulated tt̅� plott̅ed as points• Simulated background stacked• Changes to stock programs– Color changes to background for clarity– Option for non-log plots added– Color legend moved to make room for histogram

data legend

tt̅� Mass Plot

• One goal of tutorial is finding the mass of the top quark from tt̅� events– Examining decay mode producing four jets

(including two jets from b quarks), a single lepton (muon), and a neutrino

tt̅� Events

tt̅� Mass Plot

• One goal of tutorial is finding the mass of the top quark from tt̅� events– Examining decay mode producing four jets

(including two jets from b quarks), a single lepton (muon), and a neutrino

• Two ways to calculate:– Two jets from W decay plus b jet– Muon and neutrino from W decay plus b jet

Refining Plots with tt̅� Simulation

• Treated tt̅� simulated data like data in histograms– unstacked point

• Created histograms to see– what tt̅� plots should look like– how other data differed

Refining Plots with tt̅� Simulation (cont.)

• Focused on histograms of– number of jets– number of jets from b quarks– transverse momentum of jets– missing transverse energy (from neutrino)– cosine of muon angle from beamline– mass of W boson decaying to two jets

• Made trial cuts to see– how many events were cut– how cuts changed background vs. tt̅�

Interesting Preliminary Note

• Histograms of data with no cuts show W mass and t mass at approximately correct places (W ≈ 80 GeV, t ≈ 173 GeV)

Refining Plots with tt̅� Simulation (cont.)

• Cuts specific to W to 2 jets plus b jet:– Require at least 4 jets:• at least 2 b jets• at least 2 non-b jets

Refining Plots with tt̅� Simulation (cont.)

• Cuts specific to W to 2 jets plus b jet :– Require at least 4 jets:• at least 2 b jets• at least 2 non-b jets

Refining Plots with tt̅� Simulation (cont.)

• Cuts specific to W to 2 jets plus b jet :– Require mass of W decaying to 2 jets to be within

reasonable range (greater than 60 GeV, less than 110 GeV)

Refining Plots with tt̅� Simulation (cont.)

• Cuts specific to W to 2 jets plus b jet :– Require mass of W decaying to 2 jets to be within

reasonable range (greater than 60 GeV, less than 110 GeV)

Refining Plots with tt̅� Simulation (cont.)

• Other cuts– Transverse momentum of jets greater than 40

GeV

Refining Plots with tt̅� Simulation (cont.)

• Other cuts– Missing transverse energy greater than 30 GeV

Refining Plots with tt̅� Simulation (cont.)

• Other cuts– Cosine of muon angle with beamline between -

0.7 and +0.7

Refining Plots with tt̅� Simulation (cont.)

• Applying all cuts does get all tt̅� simulation and no background

Application to Data Plots

• Immediate problem!

Application to Data Plots (cont.)

• Main cause is stringent requirement on jets– Relax jet constraints to at least two non-b jets

and at least one b jet– Keep all other cuts– Note low number

of events kept

Conclusions

• Top mass becomes apparent from decay mode constraints– Other cuts are refinements

• tt̅� event rarity makes obtaining large numbers of good quality events difficult– Kept about 30 from around 400,000– Several other data sets that this analysis may be

applicable to

Into the Future

• Mass analysis for other half of tt̅� decay mode– Muon, neutrino, and b jet– MET from neutrino means more challenge

• Documentation and data archiving• Calculating cross-section• Changes to Plott̅er and/or general program

structure