Techniques and Technologies for Visualizing Scientific Data on the Web

[laurenwalker]

Organizational Page: VizSciData Category: Discussion Title: Techniques and Technologies for Visualizing Scientific Data on the Web Proposed by: Lauren Walker Participants: Summary: This session will be a collaborative sharing experience where new visualization tools and technologies are shared with the group. "New" can be something that was developed recently or an established technology that others are looking to learn.

The focus will be on web tools, technologies, and techniques for:

• visualizing scientific data
• user interfaces for discovering and investigating scientific data

The idea is that each group member will come having something to share with the group or be looking to discover new tools to use for their own projects. The group will demo these tools and discuss how they could work with our projects.

[naupaka]

To include, e.g.: ggplot, ggmap, ggvis, shiny for R, matplotlib and iPython notebooks for python, plot.ly and D3.js for the web?

[laurenwalker]

Stacy Rebich Hespanha has a topic mapping visualization project she has been wanting to implement in D3.js. I am planning on doing some coding with her on that project. This session may evolve from Discussion to Coding for just that project, depending on the amount of interest in this demo/discussion session.

[leinfelder]

Perhaps there should be another session devoted to data vis with R libraries since covering web and R technologies might be too broad for a single session?

On Jul 15, 2014, at 8:32 AM, Lauren Walker notifications@github.com wrote:

Stacy Rebich Hespanha has a topic mapping visualization project she has been wanting to implement in D3.js. I am planning on doing some coding with her on that project. This session may evolve from Discussion to Coding for just that project, depending on the amount of interest in this demo/discussion session.

— Reply to this email directly or view it on GitHub.

[laurenwalker]

Good idea, I edited the session description to focus on only web technologies. If someone has an interest in leading an R data viz session, please create a new session ticket. I'd be happy to attend, but mostly as a learner.

[bbest]

It's interesting that so many of the R packages are moving towards directly using these awesome javascript visualization libraries. For instance, the latest ggvis R library utilizes Vega which is a higher level wrapper of D3 in Node.js. Then there's libraries like the excellent rCharts (see rCharts gallery) and rMaps by ramnathv that implement a whole suite of javascript libraries, many of which are D3 derivatives. Worth also mentioning googleVis which requires internet connection to use Google Charts API . All of the libraries mentioned then also integrate with the R package Shiny for developing a web app locally (bootstrap interface to interactive R functions built from just 2 files ui.R and server.R) with the option to deploy through the web, eg ShinyApps.io.

Just noticing the awesome plotly package for ggplot2 style interactive plots.

Would be fun to work on a couple visualization examples, including Stacy's mapping visualization. I'm interested in a more interactive Ocean Health Index flower plot similar to GoogleVis' Donut chart or plotly polar area chart and visualizing linkages between OHI layers, dimensions and goals with something like rChart's Sankey diagrams.

PS The nascent ggvis package is highly promising but to generate output graphics for publication the export_png or export_svg functions currently require you to install cairo, Node and Vega to utilize the Vega command line tools vg2png and vg2svg. A common interactive capability is to see values popup on hover for a scatter plot, which I just discovered is doable with add_tooltip.

[jczaplew]

I am also interested in this session, especially with anything that involves JavaScript and spatial data. I have a good amount of experience with D3, Leaflet, and Node, so if anyone has project ideas that involve any of those technologies, I would love to get involved.

[joemudge]

I would really get a lot benefit from a demo of how to generate an interactive data visualization website from R code, using something like Shiny by RStudio, if there's time for it, interest from others, and someone willing to lead the demo.

[chrismattmann]

Take a look at Blaze and Bokeh too from Continuum Analytics. FYI @aterrel looping you in.

[karawoo]

@netthinker I'd be interested in attending a Shiny demo session too.

[chrismattmann]

Guys this will also play into the other session, issue #26 since that is prepping for an upcoming NSF Workshop on Polar Data Visualization.

[aterrel]

I can do a tutorial with Bokeh and Blaze. Have some interesting interactions with NASA data, some social graphs, and other sets.

Give me a data set to visualize beforehand and I can look at it.

[chrismattmann]

Thanks @aterrel. Check out the Arctic Data Explorer from NSDIC:

http://nsidc.org/acadis/search/

And ACADIS from NSF:

https://www.aoncadis.org/home.htm

and NASA Antarctic Master Directory (AMD):

http://gcmd.gsfc.nasa.gov/KeywordSearch/Home.do?Portal=amd&MetadataType=0

None of these datasets are clean the moment (that's one of the activities and purpose for #26 ), but if you want to take a look at a few Polar/Cyber ones those would be them. cc @abburgess

[aashish24]

I can try to present something using ParaViewWeb and GeoJS (time permitting).

[laurenwalker]

Thanks everyone for coming to this session. If you'd like to view my slides and JSFiddle on D3.js, you can view my slides on Google Drive:

https://docs.google.com/a/nceas.ucsb.edu/presentation/d/1tyc9hzbrCdA47cG6xV4YvZAQJBikZLngdLTtVkPWg_E/edit#slide=id.p

Feel free to leave comments inside the Google Doc.

[laurenwalker]

The coding topic visualization session has it's own GitHub repo here: https://github.com/lauren-walker/TopicViz

The project is totally open for questions, suggestions, and code or design contributions. The Wiki page is open for editing by anyone and we've had some discussion on the Issues page.

If you're interested in looking at the D3 visualizations that was demonstrated at the end of the codefest, you can view it online without checking out the repository: The D3 version http://lauren-walker.github.io/TopicViz/