godatadriven / taster-sessions

GDD Training Taster Sessions

Introduction

This repository contains all notebooks and materials for each GDD Training Taster.

Running a taster session

If you need to use one of these tasters, clone the repository and switch to the correct branch (see below).

If you are creating a new course, please add it as a new branch to this repo and add it to the list below.

Session Info

Each taster is run for our public courses. To access the material for the course you can go to the branch corresponding to the taster. Here is a breakdown of the branch names:

Branch Name	Taster Name	Timing	Description
pfda	Python for Data Analysts	2 hr	Public: Covers Python Essentials and Pandas with some visualisations to demonstrate the power of Python
dswp	Data Science with Python	2hr	Public: An introduction to Machine Learning followed by a demo of using sci-kit learn on the penguins dataset
adwsp	Advanced Data Science with Python	2hr	Public: An introduction to all topics covered in the ADWSP course followed by demo of feature engineering
pandas	Pandas	40m	Marysia Winkels
anomaly-detection-in-time-series	Anomaly Detection in Time Series	2hr	Public data science taster
deep-learning	Deep Learning	1hr	Public deep learning webinar
seasonality-modelling	Seasonality Modeling from Scratch	2.5hrs	PyData AMS Code Breakfast - https://youtu.be/omEVdUS14SU
neural-network-vulnerabilities	Vulnerabilities of Neural Networks: Find, Defend, & Prevent	2.5hrs	Vadim Nelidov
databricks-gcp	Machine Learning with Databricks in GCP	2hr	An intro to Databricks in GCP and following an e2e machine learning example

Contributing

To add a new taster session, create a new branch (eg. pfda) and add all the files you need there. You can branch off from taster-template to see what kind of files you need and delete all files/folders that you don't need. Each branch will need a requirements.txt file if connecting to binder (see below).

Once you have finished push the branch to GitHub but don't merge into master. Each branch is its own taster. Instead on the master branch add the information of your taster to the table above.

Using Binder

Each course can correspond to a Binder link (redirected with rebrandley) to allow participants to use jupyter notebooks during the training.

To create a binder link for a taster, make sure you have a requirements.txt file with all the packages needed for the taster. Then visit Binder, paste the URL of this repo and include the name of the branch when creating:

Contacts

Please reach out to the data science training team - Lucy Sheppard, James Hayward, Marysia Winkels, Hertbert van Leeuwen - if you have any questions.

taster-sessions's Issues

`adswp`: Feature generation for holiday is wrong

• Make sure is_holiday uses .dt.date() instead so it can compare all times of day.
• Remove the last mean encoding.

Create binder links for all our tasters

Just added the following tasters. No binders exist for them yet. Think this can be done as and when we deliver these sessions:

• anomaly detection in time series
• neural network vulnerabilities
• pandas
• deep-learning*

Some (*) I don't think are notebook based so this doesn't apply.