Welcome to the December meetup for the PyData Edinburgh group!
Our final meetup for 2018 is extra special as we have two fantastic speakers...and Mince Pies... and an opportunity to wear your treasured Christmas jumpers 😜 Pull them out of the back of the wardrobe, and anyone who manages to wear one which is in some way related to Python/Data Science might be worthy of an extra mince pie!
We'll have our regular update on the PyData world, followed by our two speakers, no lightning talks this time.
Doors open at 1815, make sure you click attend to reserve your place!
TALK 1: Python 2 is dead! Drag your old code into the modern age.
---------------------------------------------------------------
Presenter: Becky Smith
Becky is a python/django developer at Brainnwave, working mostly on their mapping platform. She's also the founder of PyLadies Edinburgh & organiser of DjangoGirls Edinburgh.
Description:
The clock is ticking on Python 2.7, with support to be dropped in January 2020. With major dependencies such as Django, NumPy and pandas moving to Python 3 only, the time has come for even big established codebases to consider upgrading. Many organisations are still postponing for various reasons; we will attempt to demonstrate that with a bit of planning and perseverance, and the assistance of some handy tools, we can embrace the future!
This session will provide a first-hand perspective on how we upgraded a large (~70,000 lines of python code) 8-year-old Django project with multiple external dependencies from Python 2.7 to Python 3.6.
TALK 2: Counting dogs & bicycles...when all the dogs are labradors & all the bikes are mountain bikes.
--------------------------------------------------------------
Presenter: David Spence
David used to build chemical plants, then he was a management consultant, then he did mergers & acquisitions & now he is trying to be a data scientist. He has just submitted his PhD thesis at the University of Sussex and is looking forward to his viva and a job!
Description:
We often use classifiers to estimate probabilities, rather than to
make actual classifications (predicting the most likely class).
For example, we want the probability of default
rather than a list of people with over 50% chance of defaulting
(hopefully, not many!), or the estimated probability of voting for
each party etc. The standard methdods for using a classifier in this
way only work if the distributions are the same in
development and production data, which isn't always the case.
In this talk, I'll discuss some alternatives, with pros and cons,
and explain how to use techniques such as Kernel Mean Matching
and Marginalised Stacked De-noising Autoencoders to make better
estimates.
LIGHTNING TALKS
==================
There's no lightning talks this month - but don't worry, they'll be back in the new year! We already have a few folk lined up for January & February, plenty space for more so please do get in touch if you'd like to do a 5 minute talk.
LOGISTICS
===========
1815: Doors open, drinks, mince pies & networking
1845: Talks start - welcome & community announcements followed by our main speakers
We'll end with pizza and more refreshments, wrapping up by 2100.
SPONSORS & REFRESHMENTS
=============================
As always, we couldn't do this without our sponsors help to provide a venue and drink & pizzas — Cathcart Associates (https://www.cathcartassociates.com), Wood Mackenzie (https://www.woodmac.com) and Canon Medical (https://research.eu.medical.canon)
CODE OF CONDUCT
====================
The PyData Code of Conduct governs this meetup. ( http://pydata.org/code-of-conduct.html ) To discuss any issues or concerns relating to the code of conduct or the behavior of anyone at a PyData meetup, please contact NumFOCUS Executive Director Leah Silen (+1 512-222-5449; leah@numfocus.org) or the group organizer.
