PyData Berlin 2015
2015-05-29..2015-05-30
Berlin, Germany
[PyData Berlin 2014, PyData Berlin 2015, PyData Berlin 2016, PyData Berlin 2017, PyData Berlin 2018]
as the organisers put it
PyData is the home for all things related to the use of Python in data management and analysis. It brings together Python enthusiasts at a novice level and includes Tutorials and corresponding talks as well as advanced talks by experts and package developers. The conference not only focuses on the application of data science tools but also on underlying algorithms and patterns. After a very successful PyData Berlin 2014 hosted at the BCC we are thrilled to announce PyData Berlin 2015, this time taking place at Betahaus Berlin.
| title | author | length |
|---|---|---|
| Python as a Framework for Analytics and Growth Hacking | Ignacio Elola | 34m |
Shows concrete pipelines: Python glues APIs (Google Analytics, Salesforce, import io) into a repeatable KPI ETL that writes rows into spreadsheets and… | ||
| Lifecycle of Web Text Mining: Scrape to Sense | Brian Carter | 28m |
Scrapes Pillreports.net with urllib2 and BeautifulSoup and stores cleaned records in MongoDB via pymongo, keeping reports and comments separate; language… | ||
| Peachbox: Agile and Accessible Big ETL Framework | Philipp Pahl | 01h 11m |
Peachbox claims to be a Python Big ETL framework that implements the Lambda Architecture end to end; the demo shows a working pipeline that ingests Amazon… | ||
| Processing Hotel Reviews with Python | Miguel Fernando Cabrera | 46m |
Pipeline processes text with Python plus Hadoop and Luigi, handling 30 million reviews crawled daily, supporting 25 languages and over 500,000 properties with… | ||
| CostCla a cost-sensitive classification library | Alejandro Correa Bahnsen | 31m |
CostCla is a Python library (scikit‑learn pandas numpy) that implements example-dependent cost matrices and cost-aware learners and claims big profit gains… | ||
| Advanced Data Storage | Thomas Pfaff | 01h 20m |
Hands-on demos use PyTables, netCDF4, numpy, pandas to create HDF5/NetCDF files, show a 100×100×100 array (1M elements, 4 bytes each) and compress it from ~8… | ||
| Indroduction to the PySpark DataFrame API | Peter Hoffmann | 45m |
PySpark demonstrates the Spark DataFrame API (introduced in Spark 1.3) as a declarative layer that compiles queries to JVM bytecode so Python data frames get… | ||
| From the Life of a Data Scientist | Felix Wick | 54m |
Claim: productive data science means shipping models into operation not slides. Core specifics: teams at Blue Yonder build predictive applications as a service… | ||
| Panel Discussion | 40m | |
Panelists report concrete outcomes not buzz: Berlin meetups hit 900 members with 150‑seat events, companies handle data from thousands to petabytes (one… | ||
| Learning to use Docker for development | Sylvain Bellemare | 01h 11m |
Builds a Dockerfile to create an image that includes matplotlib, numpy and an IPython/Jupyter notebook, runs that Jupyter notebook in a container and maps a… | ||
| What people need to be happy at work & how you can influence a diverse team environment | Lea Böhm | 32m |
Research-backed claim: only 16% truly engaged at work and happiness measurably improves outcomes — happiness raises sales 37% productivity 31% accuracy 19% —… | ||
| Lessons learned from applying PyData to our marketing organization | Jose Luis Lopez Pino | 32m |
GetYourGuide describes a PyData-stack system that automates keyword mining, campaign automation, predictive modeling, omni‑channel integration and customer… | ||
| Rewiring the Internet for Ownership with Big Data and Blockchains | Trent McConaghy | 57m |
Claim: retrofit the web for provable ownership by combining Python big‑data crawls, machine learning and the Bitcoin blockchain. Implementation crawls ~220… | ||
| Running, walking, sitting or biking? - Motion prediction with acceleration and rotation | Paul Balzer | 22m |
Smartphone and USB IMU logs at 50 Hz produce acceleration (3 axes) and rotation rate (3 axes) time series and the pipeline extracts absolute acceleration (sqrt… | ||
| Lightning Talks | 48m | |
Six lightning talks cover concrete work and results: a DialogDroids NLP platform running in the cloud only for now; a tutorial that requires Docker installed… | ||
| Smart cars of tomorrow: real-time driving patterns | Ronert Obst | 40m |
Prototype streams car data from a Bluetooth OBD II dongle through an iOS app Herbie into a Spring XD pipeline, stores raw traces in Hadoop, caches live state… | ||
| A Data Science Operationalization Framework | Alexander Kagoshima | 32m |
A small prototype framework exposes models as a REST API and stores trained artifacts in a Redis backend, implemented in Python with Flask and compatible with… | ||
| Analysing and predicting inner-city parking space occupancy | Paul Balzer | 41m |
City parking data from Dresden VAMOS is scraped into an API and analyzed at 15‑minute resolution over a year; Pandas handles feature engineering (time of day… | ||
| Measure, don't Guess - How to find out if and where to optimize | Mike Müller | 02h 33m |
The claim is: measure before you change anything and use real experiments to decide where to spend time. Examples, scripts and notebooks use cProfile… | ||
| Real-Time Monitoring of Distributed Systems | Tobias Kuhn, Nakul Selvaraj | 29m |
Trademob measures millions of time-series points from over +700 instances and spikes to 125k requests per second and builds a Python-based real-time monitoring… | ||
| Interactive Data Visualizations with Python | Christine Doig | 01h 55m |
Concrete claims: Bokeh Python generates interactive browser graphics from Python without writing JS, with three outputs (HTML file, notebook, bokeh-server)… | ||
| Scientific computing with Python: Tools for the solution of continuous problems | Claas Abert | 44m |
Shows how Python solves real large PDEs not just theory: matrices and vectors from thousands up to billions of entries, finite‑difference prototypes sped to C… | ||
| keynote | Matthew Rocklin | 49m |
Claim: a usable out-of-core multicore numpy clone exists that turns array expressions into task graphs and runs them on one machine. It uses blocked algorithms… | ||
| nothing matches | ||