{
"cells": [
{
"cell_type": "markdown",
"id": "069b7706",
"metadata": {},
"source": [
"# df105_WBosonAnalysis\n",
"The W boson mass analysis from the ATLAS Open Data release of 2020, with RDataFrame.\n",
"\n",
"This tutorial is the analysis of the W boson mass taken from the ATLAS Open Data release in 2020\n",
"(http://opendata.atlas.cern/release/2020/documentation/). The data was recorded with the ATLAS detector\n",
"during 2016 at a center-of-mass energy of 13 TeV. W bosons are produced frequently at the LHC and\n",
"are an important background to studies of Standard Model processes, for example the Higgs boson analyses.\n",
"\n",
"The analysis is translated to a RDataFrame workflow processing up to 60 GB of simulated events and data.\n",
"By default the analysis runs on a preskimmed dataset to reduce the runtime. The full dataset can be used with\n",
"the --full-dataset argument and you can also run only on a fraction of the original dataset using the argument --lumi-scale.\n",
"\n",
"See the [corresponding spec json file](https://github.com/root-project/root/blob/master/tutorials/analysis/dataframe/df105_WBosonAnalysis.json).\n",
"\n",
"\n",
"\n",
"\n",
"**Author:** Stefan Wunsch (KIT, CERN) \n",
"This notebook tutorial was automatically generated with ROOTBOOK-izer from the macro found in the ROOT repository on Tuesday, May 19, 2026 at 08:10 PM."
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "6cae8064",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:33.391495Z",
"iopub.status.busy": "2026-05-19T20:10:33.391354Z",
"iopub.status.idle": "2026-05-19T20:10:34.409819Z",
"shell.execute_reply": "2026-05-19T20:10:34.388422Z"
}
},
"outputs": [],
"source": [
"import ROOT\n",
"import sys\n",
"import json\n",
"import argparse\n",
"import os"
]
},
{
"cell_type": "markdown",
"id": "4614e7b2",
"metadata": {},
"source": [
"Argument parsing"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "dceb767b",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:34.437083Z",
"iopub.status.busy": "2026-05-19T20:10:34.436907Z",
"iopub.status.idle": "2026-05-19T20:10:34.579115Z",
"shell.execute_reply": "2026-05-19T20:10:34.570366Z"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Run on data corresponding to 0.01 fb^-1 ...\n"
]
}
],
"source": [
"parser = argparse.ArgumentParser()\n",
"parser.add_argument(\"--lumi-scale\", type=float, default=0.001,\n",
" help=\"Run only on a fraction of the total available 10 fb^-1 (only usable together with --full-dataset)\")\n",
"parser.add_argument(\"--full-dataset\", action=\"store_true\", default=False,\n",
" help=\"Use the full dataset (use --lumi-scale to run only on a fraction of it)\")\n",
"parser.add_argument(\"-b\", action=\"store_true\", default=False, help=\"Use ROOT batch mode\")\n",
"parser.add_argument(\"-t\", action=\"store_true\", default=False, help=\"Use implicit multi threading (for the full dataset only possible with --lumi-scale 1.0)\")\n",
"if 'df105_WBosonAnalysis.py' in sys.argv[0]:\n",
" # Script\n",
" args = parser.parse_args()\n",
"else:\n",
" # Notebook\n",
" args = parser.parse_args(args=[])\n",
"\n",
"if args.b: ROOT.gROOT.SetBatch(True)\n",
"if args.t: ROOT.EnableImplicitMT()\n",
"\n",
"if not args.full_dataset: lumi_scale = 0.001 # The preskimmed dataset contains only 0.01 fb^-1\n",
"else: lumi_scale = args.lumi_scale\n",
"lumi = 10064.0\n",
"print('Run on data corresponding to {:.2f} fb^-1 ...'.format(lumi * lumi_scale / 1000.0))\n",
"\n",
"if args.full_dataset: dataset_path = \"root://eospublic.cern.ch//eos/opendata/atlas/OutreachDatasets/2020-01-22\"\n",
"else: dataset_path = \"root://eospublic.cern.ch//eos/root-eos/reduced_atlas_opendata/w\""
]
},
{
"cell_type": "markdown",
"id": "2e7d3ad3",
"metadata": {},
"source": [
"Create a ROOT dataframe for each dataset\n",
"Note that we load the filenames from the external json file placed in the same folder than this script."
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "75d89d35",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:34.580936Z",
"iopub.status.busy": "2026-05-19T20:10:34.580799Z",
"iopub.status.idle": "2026-05-19T20:10:48.052326Z",
"shell.execute_reply": "2026-05-19T20:10:48.041538Z"
}
},
"outputs": [],
"source": [
"files = json.load(open(os.path.join(ROOT.gROOT.GetTutorialsDir(), \"analysis/dataframe/df105_WBosonAnalysis.json\")))\n",
"processes = files.keys()\n",
"df = {}\n",
"xsecs = {}\n",
"sumws = {}\n",
"samples = []\n",
"for p in processes:\n",
" for d in files[p]:\n",
" # Construct the dataframes\n",
" folder = d[0] # Folder name\n",
" sample = d[1] # Sample name\n",
" xsecs[sample] = d[2] # Cross-section\n",
" sumws[sample] = d[3] # Sum of weights\n",
" num_events = d[4] # Number of events\n",
" samples.append(sample)\n",
" df[sample] = ROOT.RDataFrame(\"mini\", \"{}/1lep/{}/{}.1lep.root\".format(dataset_path, folder, sample))\n",
"\n",
" # Scale down the datasets if requested\n",
" if args.full_dataset and lumi_scale < 1.0:\n",
" df[sample] = df[sample].Range(int(num_events * lumi_scale))"
]
},
{
"cell_type": "markdown",
"id": "9111f925",
"metadata": {},
"source": [
"Select events for the analysis"
]
},
{
"cell_type": "markdown",
"id": "a43f56e7",
"metadata": {},
"source": [
"Just-in-time compile custom helper function performing complex computations"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "b9509643",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:48.070557Z",
"iopub.status.busy": "2026-05-19T20:10:48.070395Z",
"iopub.status.idle": "2026-05-19T20:10:48.585245Z",
"shell.execute_reply": "2026-05-19T20:10:48.584483Z"
}
},
"outputs": [],
"source": [
"ROOT.gInterpreter.Declare(\"\"\"\n",
"bool GoodElectronOrMuon(int type, float pt, float eta, float phi, float e, float trackd0pv, float tracksigd0pv, float z0)\n",
"{\n",
" ROOT::Math::PtEtaPhiEVector p(pt / 1000.0, eta, phi, e / 1000.0);\n",
" if (abs(z0 * sin(p.theta())) > 0.5) return false;\n",
" if (type == 11 && abs(eta) < 2.46 && (abs(eta) < 1.37 || abs(eta) > 1.52)) {\n",
" if (abs(trackd0pv / tracksigd0pv) > 5) return false;\n",
" return true;\n",
" }\n",
" if (type == 13 && abs(eta) < 2.5) {\n",
" if (abs(trackd0pv / tracksigd0pv) > 3) return false;\n",
" return true;\n",
" }\n",
" return false;\n",
"}\n",
"\"\"\")\n",
"\n",
"for s in samples:\n",
" # Select events with a muon or electron trigger and with a missing transverse energy larger than 30 GeV\n",
" df[s] = df[s].Filter(\"trigE || trigM\")\\\n",
" .Filter(\"met_et > 30000\")\n",
"\n",
" # Find events with exactly one good lepton\n",
" df[s] = df[s].Define(\"good_lep\", \"lep_isTightID && lep_pt > 35000 && lep_ptcone30 / lep_pt < 0.1 && lep_etcone20 / lep_pt < 0.1\")\\\n",
" .Filter(\"ROOT::VecOps::Sum(good_lep) == 1\")\n",
"\n",
" # Apply additional cuts in case the lepton is an electron or muon\n",
" df[s] = df[s].Define(\"idx\", \"ROOT::VecOps::ArgMax(good_lep)\")\\\n",
" .Filter(\"GoodElectronOrMuon(lep_type[idx], lep_pt[idx], lep_eta[idx], lep_phi[idx], lep_E[idx], lep_trackd0pvunbiased[idx], lep_tracksigd0pvunbiased[idx], lep_z0[idx])\")"
]
},
{
"cell_type": "markdown",
"id": "b826e0c1",
"metadata": {},
"source": [
"Apply luminosity, scale factors and MC weights for simulated events"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "7e95628f",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:48.587275Z",
"iopub.status.busy": "2026-05-19T20:10:48.587137Z",
"iopub.status.idle": "2026-05-19T20:10:48.737572Z",
"shell.execute_reply": "2026-05-19T20:10:48.737035Z"
}
},
"outputs": [],
"source": [
"for s in samples:\n",
" if \"data\" in s:\n",
" df[s] = df[s].Define(\"weight\", \"1.0\")\n",
" else:\n",
" df[s] = df[s].Define(\"weight\", \"scaleFactor_ELE * scaleFactor_MUON * scaleFactor_LepTRIGGER * scaleFactor_PILEUP * mcWeight * {} / {} * {}\".format(xsecs[s], sumws[s], lumi))"
]
},
{
"cell_type": "markdown",
"id": "974dc134",
"metadata": {},
"source": [
"Compute transverse mass of the W boson using the lepton and the missing transverse energy and make a histogram"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "d34f36f7",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:48.752459Z",
"iopub.status.busy": "2026-05-19T20:10:48.752308Z",
"iopub.status.idle": "2026-05-19T20:10:49.130994Z",
"shell.execute_reply": "2026-05-19T20:10:49.121363Z"
}
},
"outputs": [],
"source": [
"ROOT.gInterpreter.Declare(\"\"\"\n",
"float ComputeTransverseMass(float met_et, float met_phi, float lep_pt, float lep_eta, float lep_phi, float lep_e)\n",
"{\n",
" ROOT::Math::PtEtaPhiEVector met(met_et, 0, met_phi, met_et);\n",
" ROOT::Math::PtEtaPhiEVector lep(lep_pt, lep_eta, lep_phi, lep_e);\n",
" return (met + lep).Mt() / 1000.0;\n",
"}\n",
"\"\"\")\n",
"\n",
"histos = {}\n",
"for s in samples:\n",
" df[s] = df[s].Define(\"mt_w\", \"ComputeTransverseMass(met_et, met_phi, lep_pt[idx], lep_eta[idx], lep_phi[idx], lep_E[idx])\")\n",
" histos[s] = df[s].Histo1D(ROOT.RDF.TH1DModel(s, \"mt_w\", 24, 60, 180), \"mt_w\", \"weight\")"
]
},
{
"cell_type": "markdown",
"id": "96c5f325",
"metadata": {},
"source": [
"Run the event loop and merge histograms of the respective processes"
]
},
{
"cell_type": "markdown",
"id": "c30b2995",
"metadata": {},
"source": [
"RunGraphs allows to run the event loops of the separate RDataFrame graphs\n",
"concurrently. This results in an improved usage of the available resources\n",
"if each separate RDataFrame can not utilize all available resources, e.g.,\n",
"because not enough data is available."
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "e4d56155",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:49.153621Z",
"iopub.status.busy": "2026-05-19T20:10:49.153454Z",
"iopub.status.idle": "2026-05-19T20:10:53.162367Z",
"shell.execute_reply": "2026-05-19T20:10:53.161764Z"
}
},
"outputs": [],
"source": [
"ROOT.RDF.RunGraphs([histos[s] for s in samples])\n",
"\n",
"def merge_histos(label):\n",
" h = None\n",
" for i, d in enumerate(files[label]):\n",
" t = histos[d[1]].GetValue()\n",
" if i == 0: h = t.Clone()\n",
" else: h.Add(t)\n",
" h.SetNameTitle(label, label)\n",
" return h\n",
"\n",
"data = merge_histos(\"data\")\n",
"wjets = merge_histos(\"wjets\")\n",
"zjets = merge_histos(\"zjets\")\n",
"ttbar = merge_histos(\"ttbar\")\n",
"diboson = merge_histos(\"diboson\")\n",
"singletop = merge_histos(\"singletop\")"
]
},
{
"cell_type": "markdown",
"id": "33667c9b",
"metadata": {},
"source": [
"Create the plot"
]
},
{
"cell_type": "markdown",
"id": "08fbca6c",
"metadata": {},
"source": [
"Set styles"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "d3600a07",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:53.165424Z",
"iopub.status.busy": "2026-05-19T20:10:53.165278Z",
"iopub.status.idle": "2026-05-19T20:10:53.273306Z",
"shell.execute_reply": "2026-05-19T20:10:53.272636Z"
}
},
"outputs": [],
"source": [
"ROOT.gROOT.SetStyle(\"ATLAS\")"
]
},
{
"cell_type": "markdown",
"id": "7ac53257",
"metadata": {},
"source": [
"Create canvas"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "53bf14cc",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:53.275042Z",
"iopub.status.busy": "2026-05-19T20:10:53.274914Z",
"iopub.status.idle": "2026-05-19T20:10:53.417460Z",
"shell.execute_reply": "2026-05-19T20:10:53.416849Z"
}
},
"outputs": [],
"source": [
"c = ROOT.TCanvas(\"c\", \"\", 600, 600)\n",
"c.SetTickx(0)\n",
"c.SetTicky(0)\n",
"c.SetLogy()"
]
},
{
"cell_type": "markdown",
"id": "b8532cb8",
"metadata": {},
"source": [
"Draw stack with MC contributions"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "51ac18e1",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:53.421213Z",
"iopub.status.busy": "2026-05-19T20:10:53.421068Z",
"iopub.status.idle": "2026-05-19T20:10:53.644851Z",
"shell.execute_reply": "2026-05-19T20:10:53.643279Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"\n",
"
\n",
"\n",
"\n",
"\n"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"stack = ROOT.THStack()\n",
"for h, color in zip(\n",
" [singletop, diboson, ttbar, zjets, wjets],\n",
" [(0.82, 0.94, 0.76), (0.76, 0.54, 0.57), (0.61, 0.6, 0.8), (0.97, 0.81, 0.41), (0.87, 0.35, 0.42)]):\n",
" h.SetLineWidth(1)\n",
" h.SetLineColor(\"black\")\n",
" h.SetFillColor(ROOT.TColor.GetColor(*color))\n",
" # From Python 3.11 onward, implicit conversion of the tuple works too:\n",
" # h.SetFillColor(color)\n",
" stack.Add(h)\n",
"stack.Draw(\"HIST\")\n",
"stack.GetXaxis().SetLabelSize(0.04)\n",
"stack.GetXaxis().SetTitleSize(0.045)\n",
"stack.GetXaxis().SetTitleOffset(1.3)\n",
"stack.GetXaxis().SetTitle(\"m_{T}^{W#rightarrow l#nu} [GeV]\")\n",
"stack.GetYaxis().SetTitle(\"Events\")\n",
"stack.GetYaxis().SetLabelSize(0.04)\n",
"stack.GetYaxis().SetTitleSize(0.045)\n",
"stack.SetMaximum(1e10 * args.lumi_scale)\n",
"stack.SetMinimum(1)"
]
},
{
"cell_type": "markdown",
"id": "d1e55050",
"metadata": {},
"source": [
"Draw data"
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "85ac5103",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:53.657264Z",
"iopub.status.busy": "2026-05-19T20:10:53.657127Z",
"iopub.status.idle": "2026-05-19T20:10:53.771981Z",
"shell.execute_reply": "2026-05-19T20:10:53.771472Z"
}
},
"outputs": [],
"source": [
"data.SetMarkerStyle(20)\n",
"data.SetMarkerSize(1.2)\n",
"data.SetLineWidth(2)\n",
"data.SetLineColor(\"black\")\n",
"data.Draw(\"E SAME\")"
]
},
{
"cell_type": "markdown",
"id": "20e5db40",
"metadata": {},
"source": [
"Add legend"
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "b80dfb47",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:53.775149Z",
"iopub.status.busy": "2026-05-19T20:10:53.775017Z",
"iopub.status.idle": "2026-05-19T20:10:53.910416Z",
"shell.execute_reply": "2026-05-19T20:10:53.909889Z"
}
},
"outputs": [],
"source": [
"legend = ROOT.TLegend(0.60, 0.65, 0.92, 0.92)\n",
"legend.SetTextFont(42)\n",
"legend.SetFillStyle(0)\n",
"legend.SetBorderSize(0)\n",
"legend.SetTextSize(0.04)\n",
"legend.SetTextAlign(32)\n",
"legend.AddEntry(data, \"Data\" ,\"lep\")\n",
"legend.AddEntry(wjets, \"W+jets\", \"f\")\n",
"legend.AddEntry(zjets, \"Z+jets\", \"f\")\n",
"legend.AddEntry(ttbar, \"t#bar{t}\", \"f\")\n",
"legend.AddEntry(diboson, \"Diboson\", \"f\")\n",
"legend.AddEntry(singletop, \"Single top\", \"f\")\n",
"legend.Draw()"
]
},
{
"cell_type": "markdown",
"id": "ff64d777",
"metadata": {},
"source": [
"Add ATLAS label"
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "9512b121",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:53.911807Z",
"iopub.status.busy": "2026-05-19T20:10:53.911682Z",
"iopub.status.idle": "2026-05-19T20:10:54.040177Z",
"shell.execute_reply": "2026-05-19T20:10:54.039447Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
""
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"text = ROOT.TLatex()\n",
"text.SetNDC()\n",
"text.SetTextFont(72)\n",
"text.SetTextSize(0.045)\n",
"text.DrawLatex(0.21, 0.86, \"ATLAS\")\n",
"text.SetTextFont(42)\n",
"text.DrawLatex(0.21 + 0.16, 0.86, \"Open Data\")\n",
"text.SetTextSize(0.04)\n",
"text.DrawLatex(0.21, 0.80, \"#sqrt{{s}} = 13 TeV, {:.2f} fb^{{-1}}\".format(lumi * args.lumi_scale / 1000.0))"
]
},
{
"cell_type": "markdown",
"id": "ac2abf2e",
"metadata": {},
"source": [
"Save the plot"
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "7e2221d2",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:54.045210Z",
"iopub.status.busy": "2026-05-19T20:10:54.045076Z",
"iopub.status.idle": "2026-05-19T20:10:54.226355Z",
"shell.execute_reply": "2026-05-19T20:10:54.225692Z"
}
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Info in : png file df105_WBosonAnalysis.png has been created\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Saved figure to df105_WBosonAnalysis.png\n"
]
}
],
"source": [
"c.SaveAs(\"df105_WBosonAnalysis.png\")\n",
"print(\"Saved figure to df105_WBosonAnalysis.png\")"
]
},
{
"cell_type": "markdown",
"id": "1d405eea",
"metadata": {},
"source": [
"Draw all canvases "
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "da72ac97",
"metadata": {
"collapsed": false,
"execution": {
"iopub.execute_input": "2026-05-19T20:10:54.233084Z",
"iopub.status.busy": "2026-05-19T20:10:54.232948Z",
"iopub.status.idle": "2026-05-19T20:10:54.356891Z",
"shell.execute_reply": "2026-05-19T20:10:54.349332Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"\n",
"
\n",
"\n",
"\n",
"\n"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from ROOT import gROOT \n",
"gROOT.GetListOfCanvases().Draw()"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.12.12"
}
},
"nbformat": 4,
"nbformat_minor": 5
}