Sunday, March 10, 2019
Workshop Session 1
12:00 - 2:00pm: A Hands-on Introduction to Deep Learning Using PyTorch
Yannet Interian, University of San Francisco
Abstract: PyTorch is a popular framework for deep learning. This is a 2-hour workshop for PyTorch beginners. We will walk you through the basics of the PyTorch library, using a text classification application as an example. During the workshop, we will use a Jupyter notebook, which will allow participants to directly experiment with the code.
12:00 - 2:00pm: An Introduction to Computational Statistics in Python
Brian Spiering, University of San Francisco
Abstract: This workshop will provide a hands-on introduction to computational statistics. The focus will be on efficiently simulating probabilities and statistics, for example calculating the outcomes of rolling dice or an A/B test. By the end of the workshop, you should be able to apply bootstrapping and permutation testing to solve applied Data Science problems.
A working knowledge of Python (e.g., creating variables and functions) is required. There are no math, probability, or statistics prerequisites.
You'll be writing code during the workshop so please bring a charged laptop. All materials and resources will be provided - https://github.com/brianspiering/ComputationalStatistics.
Workshop Session 2
2:30 - 4:30pm: Sequence to Sequence Models with Attention
Ananth Sankar, LinkedIn
Abstract: Sequence to sequence models are used to predict an output sequence given an input sequence, and can be used for applications such as query re-writing, smart reply, and machine translation. This workshop provides a hands-on introduction to sequence to sequence models that include encoder, decoder, and attention layers. We will introduce the theory of the model, and then do a hands-on exercise for a machine translation application.
We will be using PyTorch and jupyter notebooks. A working knowledge of Python is expected. Please bring fully your fully charged laptop with anaconda and PyTorch installed.
You can download the tools as follows:
Anaconda
https://www.anaconda.com/download
PyTorch
https://pytorch.org/get-started/locally/
Run the command: conda install pytorch torchvision -c pytorch
Jupyter should be installed when you install anaconda
https://jupyter.readthedocs.io/en/latest/install.html
2:30 - 4:30pm: Augmenting Data Analysis with Data from the US Census
James Livsey, US Census Bureau
This workshop will get people started and familiar with adding US Census Bureau data to their analysis. Some surveys conducted at the Census Bureau include the American Community Survey, American Housing Survey, Census of Governments, Current Population Survey, Survey of Business Owners and Survey of Income and Program Participation.
In this workshop attendees will learn about the types of data products that are freely available from such surveys; including population estimates of education, employment, family/relationships, health insurance coverage, veteran status, commuting time and thousands more.
All variables can be conjoined to existing analysis or explored on there own.
To make data accessible to the general public (geographers, city planners, etc) often it is not conveniently packaged for statisticians. Part one of the workshop will focus on the best practices to get data into a workable form for statisticians. Specifically, we focus on input and analysis in R. This discussion will include the different levels of geography available and how each relate to more classic notions of geography such as zip codes. We will discuss mapping utilities and shape files associated with the spatial data, covering robust methods that are applicable to future inquires.
The second portion of the workshop will switch to a more technical problem of changing support. More often than not data will not be available for the exact geography desired. For example, if you require housing prices based on school district the Census Bureau does not offer school district as a level of geography. Using spatial-temporal change of support techniques attendees will learn how
Workshop Session 1
12:00 - 2:00pm: A Hands-on Introduction to Deep Learning Using PyTorch
Yannet Interian, University of San Francisco
Abstract: PyTorch is a popular framework for deep learning. This is a 2-hour workshop for PyTorch beginners. We will walk you through the basics of the PyTorch library, using a text classification application as an example. During the workshop, we will use a Jupyter notebook, which will allow participants to directly experiment with the code.
12:00 - 2:00pm: An Introduction to Computational Statistics in Python
Brian Spiering, University of San Francisco
Abstract: This workshop will provide a hands-on introduction to computational statistics. The focus will be on efficiently simulating probabilities and statistics, for example calculating the outcomes of rolling dice or an A/B test. By the end of the workshop, you should be able to apply bootstrapping and permutation testing to solve applied Data Science problems.
A working knowledge of Python (e.g., creating variables and functions) is required. There are no math, probability, or statistics prerequisites.
You'll be writing code during the workshop so please bring a charged laptop. All materials and resources will be provided - https://github.com/brianspiering/ComputationalStatistics.
Workshop Session 2
2:30 - 4:30pm: Sequence to Sequence Models with Attention
Ananth Sankar, LinkedIn
Abstract: Sequence to sequence models are used to predict an output sequence given an input sequence, and can be used for applications such as query re-writing, smart reply, and machine translation. This workshop provides a hands-on introduction to sequence to sequence models that include encoder, decoder, and attention layers. We will introduce the theory of the model, and then do a hands-on exercise for a machine translation application.
We will be using PyTorch and jupyter notebooks. A working knowledge of Python is expected. Please bring fully your fully charged laptop with anaconda and PyTorch installed.
You can download the tools as follows:
Anaconda
https://www.anaconda.com/download
PyTorch
https://pytorch.org/get-started/locally/
Run the command: conda install pytorch torchvision -c pytorch
Jupyter should be installed when you install anaconda
https://jupyter.readthedocs.io/en/latest/install.html
2:30 - 4:30pm: Augmenting Data Analysis with Data from the US Census
James Livsey, US Census Bureau
This workshop will get people started and familiar with adding US Census Bureau data to their analysis. Some surveys conducted at the Census Bureau include the American Community Survey, American Housing Survey, Census of Governments, Current Population Survey, Survey of Business Owners and Survey of Income and Program Participation.
In this workshop attendees will learn about the types of data products that are freely available from such surveys; including population estimates of education, employment, family/relationships, health insurance coverage, veteran status, commuting time and thousands more.
All variables can be conjoined to existing analysis or explored on there own.
To make data accessible to the general public (geographers, city planners, etc) often it is not conveniently packaged for statisticians. Part one of the workshop will focus on the best practices to get data into a workable form for statisticians. Specifically, we focus on input and analysis in R. This discussion will include the different levels of geography available and how each relate to more classic notions of geography such as zip codes. We will discuss mapping utilities and shape files associated with the spatial data, covering robust methods that are applicable to future inquires.
The second portion of the workshop will switch to a more technical problem of changing support. More often than not data will not be available for the exact geography desired. For example, if you require housing prices based on school district the Census Bureau does not offer school district as a level of geography. Using spatial-temporal change of support techniques attendees will learn how