Requirements Document Template - UCLA Health

11.Oracle SQL Developera. Accessing Oracle SQL DeveloperTo access Oracle SQL Developer, only use the icon on the VDI Desktop. There is a version in the startmenu, but it is an old version that will not work.b. Configuring Oracle SQL DeveloperThere are a few configurations that you will need to make when you first open Oracle SQL Developer.These are very important settings, so please make sure to set them up.TNS Name File Location settingOccasionally the tns names file is updated. To make sure that you always use the latest version youwill need to set the location of the Oracle SQL Developer application looks for the tns names file.From the Tools Menu, select Preferences. Then expand Database and click on Advanced. At thebottom of the options you will see “Tnsnames Directory” browse dminSelect and click on OK to save the configuration.Semantic Analysis Info TipChanging this setting is very important. The default setting in Oracle SQL Developer can cause yourimage to crash in certain instances.From the Tools Menu, select Preferences. Then expand the Code Editor section and click onCompletion Insight. Uncheck the last option “Enable semantic analysis info tip”. Then click OK to saveyour preferences.VDI User Guide – version 2 Revised 2020092215

VDI User Guide – version 2 Revised 2020092216

12.Visual Studio 2015, 2017, and 2019a. Required licenseIn order to use Visual Studio 2015 or 2017 you must have an MSDN Account with Visual StudioProfessional license associated with it. If you purchase your license through Software Central, whenthey provide you with your license information, you will need to reply back to them to have the licenseassociated with your MSDN account. Please remember to provide them with your MSDN accountusername.Then when you open Visual Studio for the first time, remember to login using your MSDN account.b. Azure Dev OpsAlways check-in your code before exiting Visual Studio or the environment. If you have log-in issues,your profile might be reset and you could lose your changes.c. Source Control Drive MappingDue to the way that Microsoft Visual Studio saves Source Control Mapping information, it may notretain your drive mapping between sessions. You will need to re-enter your Source Control mappingeach time for a while.Basically, Microsoft stores the information relative to the machine name that you are on. When you logon to VDI you may or may not be on the same physical machine that you were on the last time. If youare, then the mapping will appear, if not, you will have to re-enter it. Once you have been on every VDIserver and made your mapping, then it will show up every time you login. The Citrix VDI randomlyassigns you to an available server so there is no way to insure that you will be logged onto the sameserver each time.VDI User Guide – version 2 Revised 2020092217

13.Appendixa. Python LibrariesThis list of libraries was last updated in May 2020Python LibraryCategoryDescriptioncore & StatisticsUse this for data manipulation and analysis. In particular, itcontains high-level data structures and tools designed for fastand easy data analysis operations. Pandas is built on NumPyand makes it easy to use in NumPy-centric applications, suchas data structures with labelled axes. Explicit data alignmentprevents common errors that result from misaligned datacoming in from different sources.It is also easy to handlemissing data using Python. Pandas is the best tool for doingdata munging.core & statisticsAllows users to explore data, estimate statistical models, andperform statistical tests. An extensive list of descriptivestatistics, statistical tests, plotting functions, and result statisticsare available for different types of data and each estimator.core & statisticsAn open source extension module for Python. The moduleNumPy provides fast precompiled functions for numericalroutines. It adds support to Python for large, multi-dimensionalarrays and matrices. Besides that it supplies a large library ofhigh-level mathematical functions to operate on these arrays.core & statisticsSciPy is widely used in scientific and technical computing.SciPy contains modules for optimization, linear algebra,integration, interpolation, special functions, FFT, signal andimage processing, ODE solvers and other tasks common inscience and engineering.data visualizationmatplotlib is a plotting library for NumPy. It provides an objectoriented API for embedding plots into applications usinggeneral-purpose GUI toolkits like wxPython, Qt, or GTK .PandasStatsmodelsNumPySciPymatplotlibVDI User Guide – version 2 Revised 2020092218

data visualizationggplot is based on ggplot2, an R plotting system. ggplotoperates differently than matplotlib: it lets you layer componentsto create a complete plot. Seasoned matplotlib users mightneed time to adjust to this new mindset. According to thecreator, ggplot isn’t designed for creating highly customizedgraphics. It sacrifices complexity for a simpler method ofplotting. ggplot is tightly integrated with pandas, so it’s best tostore your data in a DataFrame when using ggplot.data visualizationHarnesses the power of matplotlib to create beautiful charts in afew lines of code. The key difference is Seaborn’s default stylesand color palettes, which are designed to be more aestheticallypleasing and modern. Since Seaborn is built on top ofmatplotlib, you’ll need to know matplotlib to tweak Seaborn’sdefaults.data visualizationDealing with missing data is a pain. missingno allows you toquickly gauge the completeness of a dataset with a visualsummary, instead of trudging through a table. You can filter andsort data based on completion or spot correlations with aheatmap or a dendrogram.data visualizationLike ggplot, Bokeh is based on The Grammar of Graphics, butunlike ggplot, it’s native to Python, not ported over from R. Itsstrength lies in the ability to create interactive, web-ready plots,which can be easily output as JSON objects, HTML documents,or interactive web applications. Bokeh also supports streamingand real-time data.deep learningTheano is a Python library that allows you to define, optimize,and evaluate mathematical expressions involving multidimensional arrays efficiently.machine learningAn open source library for the Python. It features variousclassification, regression and clustering algorithms includingsupport vector machines, logistic regression, naive Bayes,random forests, gradient boosting, k-means and DBSCAN, andis designed to interoperate with the Python numerical andscientific libraries NumPy and SciPy.machine learningConvenient impletmentation of gradient boosting. Theselibraries provide highly optimized, scalable and fastimplementations of gradient boosting, which makes themextremely arnXGBoostVDI User Guide – version 2 Revised 2020092219

machine learningOften the results of machine learning models predictions arenot entirely clear, and this is the challenge that eli5 library helpsto deal with. It is a package for visualization and debuggingmachine learning models and tracking the work of an algorithmstep by step. It provides support for scikit-learn, XGBoost,LightGBM, lightning, and sklearn-crfsuite libraries and performsthe different tasks for each of them.machine learningA machine learning library built on top of NumPy/SciPy, theGNU Scientific Library. mlpy provides a wide range of machinelearning methods for supervised and unsupervisedproblem.mlpy is multi platform, it works with Python 2 and 3.machine learningPyBrainPyBrain is an acronym for “Python-Based ReinforcementLearning, Artificial Intelligence, and Neural Network”. It is anopen source library mainly used for neural networks,reinforcement learning and unsupervised learning. Neuralnetwork forms the basis for this library, making it a powerful toolfor real-time analytics.natural languageprocessingNLTKThe Natural Language Toolkit, or more commonly NLTK, is asuite of libraries and programs statistical natural languageprocessing (NLP) for the Python. NLTK includes graphicaldemonstrations and sample data.NLTK has been usedsuccessfully as a platform for prototyping and building researchsystems.topic modelingGensim is a ython library for topic modeling. It is built on Numpyand Scipy.Eli5MlpyGensimVDI User Guide – version 2 Revised 2020092220

b. R LibrariesThis list of libraries was last updated in May 2020ESSENTIALSAn ecosystemof useful datasciencepackagesAn opinionated collection of R packages designed for data science.All packages share an underlying design philosophy, grammar, anddata structures.readrdata importContained in the tidyverse. Base R handles most of these functions;but if you have huge files, this is a speedy and standardized way toread tabular files such as CSVs into R data frames, as well as plaintext files into character strings with read file. CRAN.readxldata importContained in the tidyverse. Fast way to read Excel files (.xls and.xlsx) in R, without dependencies such as Java. CRAN.havendata import,data wranglingContained in the tidyverse. Haven enables R to read and writevarious data formats used by other statistical packages by wrappingthe fantastic ReadStat C library written by Evan Miller. CRAN.xml2data import,data wranglingContained in the tidyverse. Work with XML files using a simple,consistent interface. Built on top of the 'libxml2' C library.httrdata import,data wranglingContained in the tidyverse. Useful tools for working with URLs andHTTP. An R interface to http protocols; useful for pulling data fromAPIs.jsonlitedata import,data wranglingContained in the tidyverse. Parse json within R or turn R dataframes into json. CRAN.rvestdata scrapingContained in the tidyverse. Easily Harvest (scrape) Web Pagesdata sharingContained in the tidyverse. Feather: fast, interoperable binary dataframe storage for Python, R, and more powered by Apache Arrow.This binary data-file format can be read by both Python and R,making data interchange easier between the two languages. It'salso built for I/O speed. CRAN.tidyversefeatherVDI User Guide – version 2 Revised 2020092221

tidyrpurrrdata wranglingdata wranglingContained in the tidyverse. While I still prefer theolder reshape2package for some general re-arranging, tidyr wonme over with specialized functions like fill (fill in missing columnsfrom data above) and replace na. Its main purpose is helping youchange data row and column formats from "wide" to "long". CRAN.Contained in the tidyverse. purrr makes it easy to apply a function toeach item in a list and return results in the format of your choice. It'smore complex to learn than the older plyr package, but also morerobust. And, its functions are more standardized than base R'sapply family -- plus it's got functions for tasks like error-checking.CRAN.data wranglingContained in the tidyverse. Tibbles are data.frames that are lazyand surly: they do less (i.e. they don’t change variable names ortypes, and don’t do partial matching) and complain more (e.g. whena variable does not exist). This forces you to confront problemsearlier, typically leading to cleaner, more expressive code. Tibblesalso have an enhanced print() method which makes them easier touse with large datasets containing complex objects.stringrdata wranglingContained in the tidyverse. Numerous functions for textmanipulation. Some are similar to existing base R functions but in amore standard format, including working with regular expressions.Some of my favorites: str pad and str trim. CRAN.forcatsdata wranglingContained in the tidyverse. Provides a suite of useful tools thatsolve common problems with factors.hmsdata wranglingContained in the tidyverse. Pretty time of Day (working with times).CRAN.lubridatedata wranglingContained in the tidyverse. Everything you ever wanted to do withdate arithmetic, although understanding & using availablefunctionality can be somewhat complex. Date-time data can befrustrating to work with in R. Lubridate makes life better. CRAN.dplyrdata wrangling,data analysisContained in the tidyverse. The essential data-munging R packagewhen working with data frames. Especially useful for operating ondata by categories. CRAN.modelrdata modellingContained in the tidyverse. Functions for modelling that help youseamlessly integrate modelling into a pipeline of data manipulationand visualisation.broomdata modellingContained in the tidyverse. Takes the messy output of built-infunctions in R, such as lm, nls, or t.test, and turns them into tidydata frames.tibbleVDI User Guide – version 2 Revised 2020092222

ggplot2magrittrdatavisualizationContained in the tidyverse. Powerful, flexible and well-thought-outdataviz package following 'grammar of graphics' syntax to createstatic graphics, but be prepared for a steep learning curve. CRAN.data wranglingContained in the tidyverse. This package gave us the % % symbolfor chaining R operations, but it's got other useful operators suchas % % for mutating a data frame in place and and . as aplaceholder for the original object being operated upon. s has a slew of functions aimed at helping you create yourown R packages, such as automatically running all example code inyour help files to make sure everything works. Requires Rtools onWindows and XCode on a Mac. On CRAN.packageinstallationIf you want to install R packages from GitHub, devtools was longthe go-to. However, it has a ton of other functions and some heftydependences. remotes is a lighter-weight alternative if all you wantis to install packages from GitHub as well as Bitbucket and someother sources. CRAN. (ghit is another option, but is GitHub-only.)packageinstallationDo you want to install a package from GitHub without typing out theGitHub user name along with the repo name? Whether becauseyou can't remember a package's GitHub owner's name, that nameis long/complex to type out, or you just want to save yourself a littletyping, this package is a handy option. Simply rungithubinstall("packagename") and the package will suggest anaccount; then you respond Y to install or n if it's the wrong one. Iteven includes fuzzy matching if you misspell a package name!data import,data exportrio has a good idea: Pull a lot of separate data-reading packagesinto one, so you just need to remember 2 functions: import andexport. llDATA IMPORTrioVDI User Guide – version 2 Revised 2020092223

datapastadata importData copy and paste: Meet reproducible research. If you've copieddata from the Web, a spreadsheet, or other sour

requirements of an individual, they will be given access to a particular VDI. Pages 3 – 6 of this document list all the software applications that come with each of these VDIs. This list is up-d