PyData Seattle 2017
2017-06-30..2017-07-02
Seattle, United States
[PyData Seattle 2017, PyData Seattle 2023, PyData Seattle 2025]
| title | author | length |
|---|---|---|
| So you want to be a Python expert? | James Powell | 01h 54m |
Live demos on Python 3.6 implement concrete behaviors: add a class with __init__, __repr__, __add__, __len__ and it prints and adds as expected; disassemble… | ||
| Implementing and Training Predictive Customer Lifetime Value Models in Python | Jean-Rene Gauthier, Ben Van Dyke | 36m |
Implementation walks through a Pareto/NBD hierarchical Bayesian model implemented in Jupyter notebooks, using PyMC (and noting Stan alternatives) and the… | ||
| Pandas, Pipelines, and Custom Transformers | Julie Michelman | 35m |
Shows that scikit-learn Pipelines strip DataFrame labels and offers a practical fix: implement pandas-friendly Pipeline and FeatureUnion and a set of custom… | ||
| Data Visualization and Exploration with Python | Stephen F Elston | 01h 51m |
Uses a 1985 car dataset (≈195 rows, mean price $13k median $10k range $5k–$45k) to show concrete recipes: load the Jupyter notebook, use Matplotlib, Pandas… | ||
| Applying the four step "Embed, Encode, Attend, Predict" framework to predict document similarity | Matthew Honnibal | 45m |
A four‑step recipe—Embed, Encode, Attend, Predict—is implemented with the Keras Deep Learning Library and custom attention layer code; word embeddings… | ||
| Interactive Data Analysis: Visualization and Beyond | Jeffrey Heer | 42m |
Argues concrete design choices: uses Vega/Vega‑Lite, D3, Protovis and Alt and shows Voyager and Wrangler (now Trifacta) prototypes that generate runnable… | ||
| JupyterLab+Real Time Collaboration | Brian E. Granger, Chris Colbert, Ian Rose | 29m |
JupyterLab brings notebooks, consoles, terminals and editors into one extensible UI and shows concrete results: a canvas CSV viewer loads and scrolls a 1.2… | ||
| Introduction to data analytics with pandas | Quentin Caudron | 01h 52m |
Pandas reads CSVs, converts types with to_numeric with errors="coerce" and datetime parsing, drops rows, then upsamples to daily with resample(...).asfreq()… | ||
| Keynote – Jake VanderPlas | Jake VanderPlas | 51m |
A concise roadmap: Python evolves from scripting to the PyData era with pandas (2009–2011) as a turning point; Conda resolves cross‑platform installs and… | ||
| Forecasting Time Series Data at scale with the TICK stack | Nathaniel Cook | 37m |
A workflow uses the TICK stack (Telegraf InfluxDB Chronograf Kapacitor) to train Prophet forecasts for ~400 GitHub repo star series, instantiate ~1,200 task… | ||
| Using Scattertext and the Python NLP Ecosystem for Text Visualization | Jason Kessler | 35m |
Open-source Python code produces interactive D3 charts embeddable in Jupyter using Scattertext with spaCy, Empath and agefromname and the demo corpus of 2012… | ||
| Chatbots Past, Present and Future | Rutu Mulkar-Mehta | 36m |
Traces chatbot claims to a few concrete pivots: Turing test era to data‑driven QA, Siri (2011) and Watson Jeopardy as milestones, and real demos like a Best… | ||
| In database Machine Learning with Python in SQL Server | Sumit Kumar | 37m |
Claims that Python runs where the data lives, using SQL Server 2017 machine learning services to train and serve models without data movement. Uses RevoScalePy… | ||
| Data processing with Apache Beam | Sourabh Bajaj | 38m |
Shows the Python SDK for Apache Beam implementing both batch and streaming pipelines that actually run on Spark Flink and Google Cloud Dataflow. Demonstrates… | ||
| Writing a Book in Jupyter Notebooks | Randall J. LeVeque | 27m |
A book is being developed as Jupyter notebooks as chapters with runnable Python, animations and sliders; the team uses nbconvert plus a custom replacement of… | ||
| Making packages and packaging "just work" | Michael Sarahan | 40m |
Shows exact fixes and working tools: a new conda-build 3 that renders Jinja templates, records compiler and variant info into a core dependency tree, and emits… | ||
| How to be a 10x Data Scientist | Stephanie Kim | 42m |
Concrete practices and measurable outcomes replace vague promises: turn models and pickled weights into scalable API endpoints or containerized microservices… | ||
| Robust Automated Forecasting in Python and R | Pranav Bahl, Jonathan Stacks | 42m |
Automated pipeline makes millions of time‑series forecasts using Python and R and wins by specifics: 2.6 million forecasts in ~5.5 hours on 28 c4.8xlarge… | ||
| Machine Learning Infrastructure at Stripe: Bridging from Python JVM | Rob Story | 32m |
Stripe tiles models trained in Python with scikit-learn and pandas and actually ships them into a Scala scoring service via a homegrown scikit-learn to Scala… | ||
| Python for NET or NET for Python | Denis Akhiyarov, Xavier Dupré | 40m |
Bridge between CPython and .NET that wraps the CPython C-API so .NET apps import numpy, scipy, pandas, matplotlib, sympy, pyomo with minimal boilerplate and… | ||
| Online Change Point Detection Using Spark Streaming | Michal Monselise | 28m |
Scan Statistics are applied to streaming data using Spark Streaming and a homemade Python package that implements Bayesian online change point detection with a… | ||
| High Fidelity Web Crawling in Python | Josh Weissbock | 29m |
Shows concrete techniques: use Python Requests for simple pages and Selenium + headless browsers (PhantomJS now moving to Chromium) for JavaScript, add logging… | ||
| D’oh! Unevenly spaced time series analysis of The Simpsons in Pandas | Joe McCarthy | 01h 48m |
Uses Pandas and the data.world API to load The Simpsons CSVs, cleans mixed-type columns, and builds a per‑character DataFrame (num_lines, num_episodes… | ||
| Bokeh and Friends | Bryan Van de Ven | 42m |
Bokeh is a Python+JS visualization stack that claims near‑1.0 stability, has 237 contributors, and now exports PNG/SVG (0.12.6). Code examples actually run: a… | ||
| bqplot Interactive Data Visualization in Jupyter | Dhruv Madeka | 31m |
bqplot is an interactive 2D plotting library for Jupyter that exposes every chart attribute as an ipywidget and links Python to D3.js/SVG in the browser… | ||
| pomegranate: fast and flexible probabilistic modeling in python | Maxwell W Libbrecht | 01h 44m |
pomegranate is a Cython package that claims concrete speed and flexibility wins: mixtures, hidden Markov models, and Bayesian networks all run 2x or more… | ||
| BrainDrain Using Machine Learning and Brain Waves to Detect Errors in Human | Katie Porterfield | 26m |
Uses the Muse Headband to record EEG, cleans CSVs, aligns timestamps, and trains Python models in Keras; experiments use raw FFT features vs a recommended… | ||
| Batch and Streaming Processing in the World of Data Engineering and Data Science | Keira Zhou | 29m |
Compares concrete tradeoffs: batch gives higher control and easier correctness, streaming gives lower latency and faster model adaptation; presents a demo that… | ||
| Robust Algorithms for Machine Learning | Tom Radcliffe | 53m |
Argues that throwing away cardinal values often helps: replace values with ranking and counting or signs to resist outliers, use SciPy stats for Mann‑Whitney… | ||
| Provenance for Reproducible Data Science | Andreas Schreiber | 34m |
Claims that provenance captured with the W3C PROV model and Python libraries can make data analytics reproducible are backed by concrete work: provenance from… | ||
| Scaling Scikit Learn | Stephen Hoover | 34m |
A scikit-learn-based backend runs training on EC2 (training data must fit memory) and pushes predictions over Redshift tables by splitting Redshift exports… | ||
| Scalable Data Science in Python and R on Apache Spark | Felix Cheung | 42m |
Concrete claims and tools: PySpark and SparkR expose the JVM-backed execution so most heavy lifting runs in the JVM and avoids language penalty; Spark ML… | ||
| Practical Optimization for Stats Nerds | Ryan J. O'Neil | 39m |
Shows how familiar estimators are reformulations for solvers and then actually builds them: least squares via closed form, via numpy, and again as a quadratic… | ||
| Python and IoT: From Chips and Bits to Data Science | Jeff Fischer | 39m |
End-to-end project that builds a smart lighting system: sensor nodes running esp8266 with MicroPython publish minute samples over MQTT to a Raspberry Pi which… | ||
| Using CNTKs Python Interface for Deep Learning | Dave DeBarr | 01h 57m |
Hands-on use of Microsoft's Cognitive Toolkit via its Python/Keras interfaces with concrete notebooks: installation steps, Azure NC6 GPU setup (Azure NC6 (K80)… | ||
| Sirbarksalot Bark Detection in Python | Nicholas Kridler | 24m |
Built app Sirbarksalot detects dachshund barks from a spare webcam, generates spectrograms, and alerts via the Facebook Messenger API. Detection starts with… | ||
| From Novice to Data Ninja | Valentina Staneva | 01h 55m |
Two‑hour hands‑on tutorial that actually builds working examples: clones the Novice2DataNinja repo, reads CSVs with pandas, computes great‑circle distances… | ||
| PixieDust make Jupyter Notebooks with Apache Spark Faster, Flexible, and Easier to use | Raj Singh | 37m |
PixieDust claims to make Jupyter + Apache Spark faster and easier by doing concrete things: automated install of Python and Scala kernels, a Scala-Python… | ||
| Parallelizing Scientific Python with Dask | Jim Crist, David Mertz | 01h 56m |
Dask claims to parallelize familiar NumPy/Pandas workflows across cores or a cluster by building task graphs and running them on schedulers. Examples show… | ||
| Make it Work, Make it Right, Make it Fast Debugging and Profiling in Dask | Jim Crist | 37m |
Shows concrete before and after: a serial pandas run ~13s reduced to ~9s then to ~4s and finally ~2s by changing algorithm and scheduler. Demonstrates tools… | ||
| A Quick Primer on TensorFrames: Apache Spark and TensorFlow Together | Denny Lee, Tomasz Drabas | 01h 56m |
TensorFrames claims to let Spark and TensorFlow exchange tensors efficiently — TensorFrames lets Spark DataFrame talk to TensorFlow without pickling by using… | ||
| High Performance Distributed Tensorflow | Chris Fregly | 42m |
Live demos apply 8-bit quantization, weight‑rounding and batch‑normalization folding to trained TensorFlow graphs, then loadtest optimized vs unoptimized… | ||
| Effective Visual Studio | Steve Dower | 02h 00m |
Visual Studio 2017 ships first‑class support for Python and R via the Data Science and Analytical Applications workload and can install Anaconda, manage… | ||
| Vocabulary Analysis of Job Descriptions | Alex Thomas | 01h 55m |
Processes a 4,300‑record corpus of US job descriptions using pandas and NLTK to tokenize, sentence‑detect, stem/lemmatize, filter stopwords and build n‑grams… | ||
| WorldRowing com: End To End Data Analysis | Lou Harwood | 37m |
WorldRowing.com data gets scraped with Python using requests and BeautifulSoup into a reproducible pipeline that required heavy cleaning because of missing IDs… | ||
| Code First, Math Later: Learning Neural Nets Through Implementation and Examples | Kyle Shaffer | 37m |
Uses Keras code to build actual models rather than theory: a convolutional net trained on 50,000 images with conv + pooling → flatten → dense → softmax, and an… | ||
| Automatic Citation generation with Natural Language Processing | Claire Kelley, Sarah Kelley | 36m |
Algorithmic citation recommendation for patents that uses topic modeling and a tf-idf / cosine similarity pipeline over 250,000 patents and abstracts, plus an… | ||
| Building a Brain Observatory for Visual Behavior | Justin Kiggins, Doug Ollerenshaw | 34m |
A Python-based pipeline integrates automated mouse training with the Allen Brain Observatory data flow, aiming to record activity across multiple cell types… | ||
| Mosaicking the Earth every day | Kelsey Jordahl | 37m |
Planet images most of the land every day with about 140 satellites after an 88‑satellite PSLV launch and now brings down 150 million square kilometers per day… | ||
| Unlocking the power of AI: A fundamentally different approach to building intelligent systems | Keen Browne | 20m |
Claims a high-level Inkling language plus an AI engine and cloud service that hide low-level network design, using a TensorFlow backend for algorithms… | ||
| Beginning Julia: Language and Landscape | en zyme | 47m |
Contains concrete demos and caveats: live REPL work with REPL, IJulia, Juno, JuliaBox shows compilation lag for packages, numeric gotchas (rational results for… | ||
| Robust Automated Forecasting in Python and R | Pranav Bahl, Jonathan Stacks | 42m |
Claims an automated pipeline that runs 2.6 million forecasts in ~5 hours by profiling and imputing daily panels, truncating incomplete seasons, detecting… | ||
| Learn to be a painter using Neural Style Painting | Pramit Choudhary | 40m |
Uses VGG-19 to extract content and style features and optimizes a loss made of content loss style loss and total variation denoising, tuning the alpha/beta… | ||
| Designing for Guidance in Machine Learning | Olivia Gunton | 43m |
Real examples and numbers: Python code uses MNIST handwritten digits, TensorFlow yields higher raw accuracy but is a black box, while engineered features plus… | ||
| Moving notebooks into the cloud: challenges and lessons learned | Saranga Komanduri, Lori Eich | 34m |
Civis Analytics claims a working cloud Jupyter integration that runs notebooks in Docker on Kubernetes, fronts them with an auth proxy issuing 10‑second… | ||
| Building a community fountain around your data stream | Maria Patterson | 43m |
Uses Apache Kafka for replicated, rewindable delivery, Avro schemas for compact, evolvable messages, and Spark streaming to run batchlike filters from Jupyter… | ||
| We came, we saw, we hacked How to win a Big Data hackathon | Eloisa Tran | 34m |
Description
Hackathon prize money keeps rising each year as organizations learn to take advantage of their value. Runner-up cash prizes are now in the… | ||
| Monitoring Displacement Crises with Python | George Richardson | 31m |
Python backend that scrapes web pages and runs an NLP pipeline that extracts numbers of people displaced and timestamps, using newspaper for parsing, gensim… | ||
| Upgrading Legacy Projects: Lessons Learned | Matt Braymer-Hayes, Erin Haswell | 35m |
Shows concrete tactics: build comparison tests against historical outputs to detect regressions, run CI via GitLab pipelines for every push, use Python -W to… | ||
| Keynote – Joseph Sirosh | Joseph Sirosh | 55m |
Shows concrete pipelines and numbers: uses MMLSpark and transfer learning on ResNet to raise snow‑leopard detection from ~68% to ~89% accuracy, runs… | ||
| Changing the Bureaucrat’s Mind Toward Data Driven Decision | Gary Dunow | 31m |
Claims that NGA is shifting culture and sharing data are supported by concrete steps: 2014 release of unclassified data on the web, investment in… | ||
| How diversity drives excellence in our data driven tech world | sofi a | 54m |
Panelists argue that representation changes outcomes, citing machine learning facial recognition bias as a concrete failure and using an ensemble models… | ||
| nothing matches | ||