Hellow guys, Welcome to my website, and you are watching [34] Taking the Edge Off of Data Science with dabl (Andreas Mueller). and this vIdeo is uploaded by Data Umbrella at 2021-10-05T13:34:18-07:00. We are pramote this video only for entertainment and educational perpose only. So, I hop you like our website.
| Name | [34] Taking the Edge Off of Data Science with dabl (Andreas Mueller) |
| Video Uploader | Video From Data Umbrella |
| Upload Date | This Video Uploaded At 05-10-2021 20:34:18 |
| Video Discription | ## Upcoming Events Join our Meetup group for more events! https://www.meetup.com/data-umbrella [34] Andreas Mueller: Taking the Edge Off of Data Science with dabl Exploratory Data Analysis ## Key Links - Transcript: https://github.com/data-umbrella/event-transcripts/blob/main/2021/34-andreas-dabl.md - Video: https://youtu.be/h92RMJi4mRM - Meetup Event: https://www.meetup.com/data-umbrella/events/280283624/ - GitHub repo: https://github.com/dabl/dabl ## Resources - lux: https://github.com/lux-org/lux/ - alt text discussion: https://twitter.com/scikit_learn/status/144541067 - Auto ML paper: https://ml.informatik.uni-freiburg.de/wp-content/uploads/papers/18-AUTOML-AutoChallenge.pdf - scikit-learn release highlights: https://scikit-learn.org/stable/whats_new/v1.0.html ## Agenda 00:00:00 Reshama introduces Data Umbrella 00:04:34 Andreas begins presentation 00:05:45 Discuss: scikit-learn 1.0 release highlights! 00:06:29 1.0: Keyword and positional arguments 00:08:06 1.0: Enforce column name consistency 00:10:52 1.0: Feature name support via pandas 00:11:30 1.0: Feature name support 00:15:07 1.0: Histogram gradient boosting is stable 00:16:30 1.0: More (splines, more plotting, quantile regression, online SVM) 00:17:55 Back to dabl; A real world ML workflow 00:21:28 Machine learning with scikit-learn and pandas 00:22:52 Example code: training a linear model in scikit-learn with pipeline (scale, impute, one-hot encode, continuous & categorical, logistic regression, grid search) 00:24:25 Automatic machine learning frameworks 00:26:35 Data Analysis Baseline Library (D-A-B-L) 00:30:00 Viz: dabl.clean, dabl.plot, dabl.AnyClassifier, dabl.explain 00:30:10 Data cleaning and preprocessing 00:33:35 Preprocessing 00:34:26 Visualization 00:38:50 Model building 00:40:45 Automatic model search, Portfolio creation 00:47:00 Mmodel explanation 00:47:55 Metrics (accuracy, precision, recall, F1, support) 00:49:17 Q: What is permutation importance? 00:52:16 Upcoming in next version of dabl (alluvial (Sankey) plot, ) 00:55:48 Future goals (time series, text data, model compression, building explainable models) 00:56:27 pip install dabl 00:57:05 Q: How can people contribute to dabl? 00:59:47 Q: Have you thought about making the plots interactive? (sharing about lux) 01:03:06 Q: What is your opinion on using permutation importances vs shapely regression vs relative weights importance? 01:06:05 Q: Do alluvial plots use causal modeling to detect relationships between predictors? 01:07:18 Q: Is dabl something that could have been part of scikit-learn? 01:10:55 Q: How does the Lux library complement dabl? 01:13:10 Q: Is interpreting models a next step in data science tools? 01:15:28 Q: Adding alt-text to visualizations. How do we describe visualizations? 01:22:07 Closing words by Andreas ## Event In many real-world applications, data quality and curation and domain knowledge play a much larger role in building successful models than coming up with complex processing techniques and tweaking hyper-parameters. Therefore, a data toolbox should enable users to understand both data and model, and not burden the practitioner with details of formatting plots, rote preprocessing steps and hyperparameters. The dabl library is a first step in this direction. It provides automatic visualization routines and model inspection capabilities while automating away model selection. This talk will introduce the dabl library and show how to use it to quickly create supervised models and identify modeling and data quality issues. ## Speaker Andreas Müller is a Principal Research SDE at Microsoft, where he works on the interface of the Data Science ecosystem and cloud infrastructure. He previously held positions as Associate Research Scientist at the Columbia Data Science Institute and as a Research Engineer at the NYU Center for Data Science. He is one of the core developers of the scikit-learn machine learning library, a member of the scikit-learn technical committee, and the author of the book "Introduction to machine learning with Python". His work focuses on practical aspects of machine learning and the development of user-centric machine learning software. LinkedIn: https://www.linkedin.com/in/andreas-mueller-b370265a/ Twitter: https://twitter.com/amuellerml GitHub: https://twitter.com/amueller #eda #python #datascience |
| Category | Science & Technology |
| Tags | python | data science | dataviz | open source |
