Share
Embed
Embracing Open Data Science in Your Organization
ACM Webinar, September 7th 2016
Christine Doig
September 07, 2016
Transcript
50 slides
-
Embracing Open Data Science in your Organization Christine Doig Senior
Data Scientist Continuum Analytics -
2 • Introduction to Data Science • Data Science Challenges
in Organizations • Anaconda Distribution • Anaconda Community Innovation • Anaconda Enterprise Platform Agenda -
5 The Data Science Venn Diagram Revisited Machine Learning Big
Data Visualization Analytics HPC CS / Programming DS -
6 Machine Learning Big Data Visualization Analytics HPC CS /
Programming DS Data Scientist come with different skills and backgrounds Machine Learning Big Data Visualization Analytics HPC CS / Programming DS Machine Learning Big Data Visualization Analytics HPC CS / Programming DS Statistician / Analyst Research / Computational Scientist Developer / Engineer -
7 Data Science in summary: • is a team sport
• formed by team members with very diverse backgrounds • both in terms of knowledge (CS, Statistics, Viz, ML…) • and technology stacks (R, SAS, Python…) How can companies organize efficiently in this environment? -
© 2016 Continuum Analytics- Confidential & Proprietary With an inclusive
movement that makes open source tools for data science -- data, analytics, & computation – easily work together as a connected ecosystem 8 Open Data Science -
© 2016 Continuum Analytics- Confidential & Proprietary Open Data Science
Vibrant and Growing Community 9 Python Community 30M+ ANACONDA Downloads* 3M+ Packages in Anaconda 720+ R Community 16M+ Spark Python Usage 50%+ * As of Dec 2015. Another 2.7M download YTD -
© 2016 Continuum Analytics- Confidential & Proprietary Availability Innovation Interoperability
Transparency For everyone in the data science team 10 Open Data Science means… OPEN DATA SCIENCE is the FOUNDATION TO MODERNIZATION -
© 2016 Continuum Analytics- Confidential & Proprietary 11 Data Scientist
Biz Analyst Data Engineer Developer DevOps Deploy & Operate Explore & Analyze Collaborate & Publish Data Scientists are not the only player in the Data Science Team -
12 Data Science assets Data Scientist Biz Analyst Developer Spreadsheets
Reports Presentations Notebooks Scripts Visualizations Software packages Web applications -
14 Data Science workflows Explore & Analyze Data Query Visualize
Clean & Tidy Predict, Simulate, & Optimize Interactive Reports Interactive Presentations Interactive Notebooks Interactive Apps Predictive Models Collaborate & Publish Interactive Notebooks Predictive Models Interactive Apps Code Applications -
15 Data Science workflows Deploy & Operate Querying & Reports
Web Services Data Warehouse HDFS Streaming Data Flat Files NoSQL Model Building Integrate DEPLOY OPERATE Cloud Computing Web Services On-Premise Internal Cluster -
17 Challenges • Manage reproducible heterogeneous Data Science environments •
Distribute, share and publish Data Science assets • Get diverse data scientists (languages, tools, data models, assets…) to collaborate effectively • Enable Data Scientists to easily leverage Big Data technologies • Deploy data science assets into production applications • Share insights with decision makers • Enable Business Analysts and Managers to leverage Data Science -
18 How are we solving those challenges through: • Anaconda
Distribution • Anaconda Community Innovation • Jupyter, JupyterLab and extensions • Bokeh for interactive data visualizations • Datashader for large scale visualizations • Dask for parallel computing • Numba for high performance computing • Anaconda Enterprise -
20 The Distribution for Data Science Machine Learning Big Data
Visualization Analytics Scientific computing CS / Programming Numba Blaze Bokeh -
23 Anaconda Distribution Glossary PYTHON NumPy, SciPy, Pandas, Scikit-learn, Jupyter
/ IPython, Numba, Matplotlib, Spyder, Numexpr, Cython, Theano, Scikit-image, NLTK, NetworkX and 150+ packages conda PYTHON cond conda • Anaconda distribution: Python distribution that includes 150+ packages for data science (in the installer) • Miniconda: Lightweight version of Anaconda, with just Python and conda. • Anaconda Cloud: Cloud service to host and share public (free) and private data science assets • Anaconda Navigator: Anaconda distribution UI to manage environments, launch applications and learn about what’s happening in the community Anaconda distribution Miniconda -
24 Anaconda Navigator Launch applications Learn about the Anaconda community
Manage environments -
25 • conda: Cross-platform and language agnostic package and environment
manager • conda-forge: A community led collection of recipes, build infrastructure and distributions for the conda package manager • conda environments: custom isolated sandboxes to easily reproduce and share data science projects • conda kapsel: reproducible, executable project directories -
26 $ conda install python=2.7 $ conda install pandas $
conda install -c r r $ conda install -c conda-forge tensorflow name: myenv channels: -chdoig -r -foo dependecies: -python=2.7 -r -r-ldavis -pandas -mongodb -spark=1.5 -pip -pip: - flask-migrate - bar=1.4 $ conda env create $ source activate myenv Install dependencies Manage multiple environments $ conda kapsel run plot --show Deploy an interactive visualization -
27 What challenges does Anaconda Distribution solve? PYTHON NumPy, SciPy,
Pandas, Scikit-learn, Jupyter / IPython, Numba, Matplotlib, Spyder, Numexpr, Cython, Theano, Scikit-image, NLTK, NetworkX and 150+ packages conda PYTHON cond conda Anaconda distribution Miniconda • Easy to install on all platforms • Language agnostic - Python, R, Scala… • Trusted by industry leaders • Trusted by the community - Large user base: 3M+ downloads • BSD license • Extensible - easily build, share and install proprietary libraries with Anaconda Cloud • Allows isolated custom sandboxes with different versions of packages - conda environments • Allows for easy encapsulation and deployment of data science assets - conda kapsel -
29 • Anaconda Distribution • Anaconda Community Innovation • Jupyter,
JupyterLab and extensions • Bokeh for interactive data visualizations • Datashader for large scale visualizations • Dask for parallel computing • Anaconda Enterprise -
30 Continuum Analytics contributions to the Python ODS ecosystem Bokeh
Dask Datashader • Web interactive data visualizations (no JS) • Graphics pipeline system for creating meaningful representations of large amounts of data • Parallel computing framework • Next generation Data Science IDE JupyterLab -
31 Jupyter Notebook Web application that allows you to create
and share documents that contain live code, equations, visualizations and explanatory text. $ jupyter notebook -
33 Sharing insights with decision makers From text, code and
visualizations directly to slides -
34 Jupyter: Extensions - nbpresent remix your Jupyter Notebooks as
interactive slideshows with a UI editor • Edit slides, layout and themes conda install -c anaconda-nb-extensions nbpresent jupyter notebook -
35 Jupyter extensions - anaconda-nb-extensions • nb_condakernel: use the kernel-switching
dropdown inside notebook UI to switch between conda envs • nb_conda: help manage conda envs from inside file viewer of jupter notebook nb_condakernel nb_conda -
36 Jupyter: IRkernel https://www.continuum.io/blog/developer/jupyter-and-conda-r conda config --add channels r conda
install r-essentials jupyter notebook Trivial to get started writing R notebooks the same way you write Python ones. -
37 Bokeh Interactive visualization framework that targets modern web browsers
for presentation • No JavaScript • Python, R, Scala and Lua bindings • Easy to embed in web applications • Server apps: data can be updated, and UI and selection events can be processed to trigger more visual updates. http://bokeh.pydata.org/en/latest/ -
39 Datashader graphics pipeline system for creating meaningful representations of
large amounts of data • Provides automatic, nearly parameter-free visualization of datasets • Allows extensive customization of each step in the data-processing pipeline • Supports automatic downsampling and re- rendering with Bokeh and the Jupyter notebook • Works well with dask and numba to handle very large datasets in and out of core (with examples using billions of datapoints) https://github.com/bokeh/datashader NYC census data by race -
41 Dask: Scaling Data Analysis Client Machine Compute Node Compute
Node Compute Node Head Node One month CSV file ~ 2GBs Two years CSV files ~ 50GB Scaling Data Analysis Six month CSV file ~ 12GBs Client Machine Compute Node Compute Node Compute Node Head Node Client Machine Compute Node Compute Node Compute Node Head Node HDFS + + distributed -
42 Dask Dataframes >>> import pandas as pd >>> df
= pd.read_csv('iris.csv') >>> df.head() sepal_length sepal_width petal_length petal_width species 0 5.1 3.5 1.4 0.2 Iris-setosa 1 4.9 3.0 1.4 0.2 Iris-setosa 2 4.7 3.2 1.3 0.2 Iris-setosa 3 4.6 3.1 1.5 0.2 Iris-setosa 4 5.0 3.6 1.4 0.2 Iris-setosa >>> max_sepal_length_setosa = df[df.species == 'setosa'].sepal_length.max() 5.7999999999999998 >>> import dask.dataframe as dd >>> ddf = dd.read_csv('*.csv') >>> ddf.head() sepal_length sepal_width petal_length petal_width species 0 5.1 3.5 1.4 0.2 Iris-setosa 1 4.9 3.0 1.4 0.2 Iris-setosa 2 4.7 3.2 1.3 0.2 Iris-setosa 3 4.6 3.1 1.5 0.2 Iris-setosa 4 5.0 3.6 1.4 0.2 Iris-setosa … >>> d_max_sepal_length_setosa = ddf[ddf.species == 'setosa'].sepal_length.max() >>> d_max_sepal_length_setosa.compute() 5.7999999999999998 Dask dataframes look and feel like pandas dataframes, but operate on datasets larger than memory using multiple threads -
43 Distributed http://distributed.readthedocs.io/en/latest/ Distributed is a lightweight library for distributed
computing in Python. It extends dask APIs to moderate sized clusters. -
44 Web UI Dask.distributed includes a web interface to help
deliver information about the current state of the network helps to track progress, identify performance issues, and debug failures over a normal web page in real time. -
© 2016 Continuum Analytics- Confidential & Proprietary ANACONDA platform 46
ANACONDA Repository ANACONDA Accelerate ANACONDA Distribution ANACONDA Scale Open Data Science Core Open Data Science Repository High Performance Computing Distributed Computing ANACONDA Enterprise Notebooks Data Science Collaboration ANACONDA Mosaic Heterogeneous Data Exploration ANACONDA Fusion Excel Data Science -
48 Challenges revisited • Manage reproducible Data Science environments •
Distribute Data Science assets • Get diverse data scientists (languages, tools, data models, deliverables…) to collaborate effectively • Enable Data Scientists to easily leverage Big Data technologies • Deploy data science assets into production applications • Share insights with decision makers -
49 https://www.continuum.io/ Learn more • Whitepapers: https://www.continuum.io/whitepapers • Webinars: https://www.continuum.io/webinars
• Presentations: https://www.continuum.io/presentations • Videos: https://www.continuum.io/videos