First draft of data analysis environments #13
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
LittleAprilFool
changed the title
cyrus-
requested changes
Nov 21, 2019
There was a problem hiding this comment.
April, great job covering a lot of different topics. This is very well organized. Most of my comments have to do with wording changes and organizational things (in particular, each paper should have a summary). In addition, you should add a bit more detail about how the various tools you discuss were evaluated.
src/index.rst
Outdated
| :caption: Data Analysis Environments | ||
| :hidden: | ||
|
|
||
| data-analysis-environments.rst |
There was a problem hiding this comment.
Let's put Data Analysis Environments under the Live Programming chapter, rather in its own top-level chapter
src/data-analysis-environments.rst
Outdated
|
|
||
| Overview | ||
| ======== | ||
| With the notion of extracting knowledge and insights from data, the term data science was first distinguished from pure statistics in 1997. |
There was a problem hiding this comment.
Before defining "data science", you should introduce "data analysis", since that is the title of the section. Something like "Data scientists use data analysis environments for a variety of data science activities. The term data science was first..."
src/data-analysis-environments.rst
Outdated
| Overview | ||
| ======== | ||
| With the notion of extracting knowledge and insights from data, the term data science was first distinguished from pure statistics in 1997. | ||
| University of Michigan statistics professor C.F. Jeff Wu popularized the term "data science" in his talk "Statistics = Data Science?" and he identified 3 aspects to data science which differentiate it from pure statistics :cite:`donoho:2017`: 1) data collection, 2) data modeling and analysis, 3) problem solving and decision support. |
There was a problem hiding this comment.
(Do statisticians not do "data modeling and analysis"?)
src/data-analysis-environments.rst
Outdated
|
|
||
| This article reflects on the past 50 years of history of data science. It introduces how the field of data science emerged as a superset of statistics and machine learning, driven by commercial rather than intellectual developments. It also visions how the field of data science will grow in the next 50 years. | ||
|
|
||
| Data science has grown rapidly over the last decade with the rise of big data. |
There was a problem hiding this comment.
"big data" is kind of a meaningless buzzword at this point, so let's avoid that and instead say something like "as datasets have increased rapidly in size and ubiquity".
src/data-analysis-environments.rst
Outdated
| This article reflects on the past 50 years of history of data science. It introduces how the field of data science emerged as a superset of statistics and machine learning, driven by commercial rather than intellectual developments. It also visions how the field of data science will grow in the next 50 years. | ||
|
|
||
| Data science has grown rapidly over the last decade with the rise of big data. | ||
| It never comes to a consensus on the workflow of data science. |
There was a problem hiding this comment.
awkward sentence, perhaps just remove it entirely.
src/data-analysis-environments.rst
Outdated
| Statistical analysis is the process to generate mathematically rigorous evaluations from data. | ||
| There are many statistical tests designed for different contexts and purposes, which may stand only under specific preconditions. | ||
| Thus, it is a difficult task for data science workers, especially people with little or no statistical expertise, to decide which statistical tests to use given a specific dataset and hypotheses. | ||
| Tea is a high-level declarative language to translate users' hypotheses and domain knowledge into all valid statistical tests :cite:`jun:2019`. |
src/data-analysis-environments.rst
Outdated
| There are many computational notebook platforms designed for different analysis languages and environments, for example, `Apache Zeppelin`_, `Spark Notebook`_, `Observable`_, `RStudio`_, `Wolfram Notebooks`_. | ||
| Among these computational notebook platforms, `Jupyter Notebook`_ is the most widely used one. | ||
| It evolved from IPython, which is a terminal-based interactive shell for creating interactive visualizations for scientific computing. | ||
| Wrapping IPython as the kernel, Jupyter Notebook has a powerful graphical interface that allows users to edit and execute "cells" -- small chunks of code or markdown text. |
There was a problem hiding this comment.
Mention that Jupyter now supports more than just Python.
src/data-analysis-environments.rst
Outdated
| .. bibliography:: data-analysis-environments2.bib | ||
| :filter: key == 'rule:2018' | ||
|
|
||
| Kery et al. took an interview approach to study how data scientists kept track of variants they explored in Jupyter notebook :cite:`kery:2018`. |
src/data-analysis-environments.rst
Outdated
| To address the challenges in informal versioning, they designed Variolite, a code editing tool with local versioning control :cite:`kery:2017`. | ||
| Variolite is an Atom editor extension that enables users to version a section of the code based on users' selection. | ||
| They later integrated this design into Jupyter notebook with Verdant :cite:`kery:2019`. | ||
| They designed an enhanced history view with algorithmic and visualization techniques for data science workers to better foraging past analysis choices. |
src/data-analysis-environments.rst
Outdated
| .. bibliography:: data-analysis-environments2.bib | ||
| :filter: key == 'kery:2019' | ||
|
|
||
| Head et al. took a different design approach :cite:`head:2019`. |
There was a problem hiding this comment.
Paper summary + talk about evaluation
cyrus-
requested changes
Dec 6, 2019
There was a problem hiding this comment.
Looks great -- main issue is that you are describing the papers in the text rather than in the corresponding paper summary. Please reorganize so that the details of specific papers are in the paper summaries (it's okay if the paper summaries are long). Second draft grade will be sent by email.
src/data-analysis-environments.bib
Outdated
|
|
||
| @inproceedings{zhang:2017, | ||
| author = {Zhang, Xiong and Guo, Philip J.}, | ||
| title = {DS.Js: Turn Any Webpage into an Example-Centric Live Programming Environment for Learning Data Science}, |
There was a problem hiding this comment.
put abbreviations in titles in braces, e.g. {DS.js}
src/data-analysis-environments.bib
Outdated
| @inproceedings{zhang:2019, | ||
| author = {Zhang, Xiong and Guo, Philip J.}, | ||
| title = {Mallard\&\#58; Turn the Web into a Contextualized Prototyping Environment for Machine Learning}, | ||
| booktitle = {Proceedings of the 32Nd Annual ACM Symposium on User Interface Software and Technology}, |
src/data-analysis-environments2.bib
Outdated
|
|
||
| @inproceedings{xia:2018, | ||
| author = {Xia, Haijun and Henry Riche, Nathalie and Chevalier, Fanny and De Araujo, Bruno and Wigdor, Daniel}, | ||
| title = {DataInk: Direct and Creative Data-Oriented Drawing}, |
src/live-programming.rst
Outdated
| @@ -16,6 +16,379 @@ REPLs | |||
|
|
|||
| Data Analysis Environments | |||
| ========================== | |||
| Data scientists use data analysis environments for a variety of data science activities. | |||
| With the notion of extracting knowledge and insights from data, the term data science was first distinguished from pure statistics in 1997 :cite:`donoho:2017`. | |||
There was a problem hiding this comment.
Start this sentence with "The term data science was first distinguished from pure statistics in 1997"
src/live-programming.rst
Outdated
| @@ -16,6 +16,379 @@ REPLs | |||
|
|
|||
| Data Analysis Environments | |||
| ========================== | |||
| Data scientists use data analysis environments for a variety of data science activities. | |||
| With the notion of extracting knowledge and insights from data, the term data science was first distinguished from pure statistics in 1997 :cite:`donoho:2017`. | |||
| University of Michigan statistics professor C.F. Jeff Wu popularized the term "data science" in his talk "Statistics = Data Science?". | |||
There was a problem hiding this comment.
no need to call out University of Michigan here
There was a problem hiding this comment.
is there a link to the talk you can include?
src/live-programming.rst
Outdated
| It helps data science workers to present, reproduce, share, and collaborate their analysis. | ||
| There are many computational notebook platforms designed for different analysis languages and environments, for example, `Apache Zeppelin`_, `Spark Notebook`_, `Observable`_, `RStudio`_, `Wolfram Notebooks`_. | ||
| Among these computational notebook platforms, `Jupyter Notebook`_ supports more than 40 programming languages and has been widely used for writing and sharing computational narratives in various contexts. | ||
| It evolved from IPython, which is a terminal-based interactive shell for creating interactive visualizations for scientific computing. |
There was a problem hiding this comment.
see if you can find a paper about IPython to cite
src/live-programming.rst
Outdated
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | ||
| Although computational notebooks are designed to support not only performing, but also documenting and sharing analysis, most people consider it personal, exploratory, and messy. | ||
|
|
||
| Rule et al. :cite:`rule:2018` analyzed over 1 million notebooks on GitHub and found them often lack explanatory text. |
src/live-programming.rst
Outdated
| This paper reports a large scale analysis of over 1 million open-source computational notebooks. | ||
| The results show that only one in four held explanatory text. | ||
|
|
||
| Kery et al. took an interview approach to study how data scientists kept track of variants they explored in Jupyter notebook :cite:`kery:2018`. |
| To address the challenges in informal versioning, they designed Variolite, a code editing tool with local versioning control :cite:`kery:2017`. | ||
| Variolite is an Atom editor extension that enables users to version a section of the code based on users' selection. | ||
| A preliminary usability study shows that 9 out of 10 participants found the tool easy to use and all 10 of them would consider use it in real life. | ||
| They later integrated this design into Jupyter notebook with Verdant :cite:`kery:2019`. |
src/live-programming.rst
Outdated
|
|
||
| This paper explores the design space in notebook code enviroments to help data scientists forage for information in their history. | ||
|
|
||
| Head et al. took a different design approach :cite:`head:2019`. |
cyrus-
approved these changes
Dec 23, 2019
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Sorry, I couldn't finish the whole draft on time. I will keep adding to this pr as I make more progress.