python -m pip install --user numpy scipy matplotlib ipython jupyter pandas sympy nose. We recommend using an user install, sending the --user flag to pip.SciPy.orgInstallation¶Installationsmethodsinclude:DistributionspipPackageManagerSourceBinariesMethodsdifferineaseofuse,coverage,maintenanceofoldversions,system-wideversuslocalenvironmentuse,andcontrol.WithpiporAnaconda’sconda,youcancontrolthepackageversionsforaspecificprojecttopreventconflicts.Condaalsocontrolsnon-Pythonpackages,likeMKLorHDF5.Systempackagemanagers,likeapt-get,installacrosstheentirecomputer,oftenhaveolderversions,anddon’thaveasmanyavailableversions.Sourcecompilationismuchmoredifficultbutisnecessaryfordebugginganddevelopment.Ifyoudon’tknowwhichinstallationmethodyouneedorprefer,werecommendtheScientificPythonDistributionAnaconda.ScientificPythonDistributions(recommended)¶Pythondistributionsprovidethelanguageitself,alongwiththemostcommonlyusedpackagesandtools.Thesedownloadablefilesrequirelittleconfiguration,workonalmostallsetups,andprovideallthecommonlyusedscientificpythontools.AnacondaworksonWindows,Mac,andLinux,providesover1,500Python/Rpackages,andisusedbyover15millionpeople.Anacondaisbestsuitedtobeginningusers;itprovidesalargecollectionoflibrariesallinone.Formoreadvanceduserswhowillneedtoinstallorupgraderegularly,Minicondaisamoresuitablewaytoinstallthecondapackagemanager.Otheroptionsinclude:WinPython:AnotherfreedistributionincludingscientificpackagesandtheSpyderIDE;Windowsonly,butmoreactivelymaintainedandsupportsthelatestPython3versions.Pyzo:AfreedistributionbasedonAnacondaandtheIEPinteractivedevelopmentenvironment;SupportsLinux,Windows,andMac.Installingviapip¶Pythoncomeswithaninbuiltpackagemanagementsystem,pip.Pipcaninstall,update,ordeleteanyofficialpackage.Youcaninstallpackagesviathecommandlinebyentering:python-mpipinstall--usernumpyscipymatplotlibipythonjupyterpandassympynoseWerecommendusinganuserinstall,sendingthe--userflagtopip.pipinstallspackagesforthelocaluseranddoesnotwritetothesystemdirector