visfest / topics2015 Goto Github PK

Any real world examples out there? Isomorphic Javascript sounds appealing. Is it? How well does D3 work with Meteor?

Possible outcomes

• Simple creation and demo
• Data considerations with MongoDB or integrating alternative data providers
• Best practices.

Some layouts & plugins are built into D3. You can also write your own, or use those written by others. Let's talk about what makes a good layout and how to create them.

Possible goal: List of some great layouts (including those not in core D3), and best practices for creating them yourself.

Quantified self is big now. But a lot of data from individuals are mostly about health data from sensing devices including mobile phone.

• What would be interesting topics for personal data visualization other than the health/fitness/wellness area?
• How can we design visualization that helps people understand themselves better than mere numbers?

I can talk about my recent project - visualization of personal beer taste and drinking behavior: that you can see here: http://beer.tany.kim/#/_sample1

Large data sets are a challenge. What can we do to intelligently aggregate data to only show what the user needs while still giving hints or cues as to what lies underneath.

Possible Outcomes:

• Discussion of Hex binning for data aggregation
• Novel approaches to hide and show the extended data linkages (graph and force layouts)
• Using a data store to swap out different subsets of a large data set.
• Performance concerns.

Basically a Tableau like tool that attracts designers to be able to quickly get common vizualizations out of their data.

This will really promote the d3 among non-programmer folks such as designers, PMs and business folks.

Not sure if there are any tools like that. Lets discuss around this.

Develop a vis that tells a story on its own but then can be easily added to for new stories.

Example: Financials of Ten Leading Tech Companies

Base Financials

This vis can stand alone as a base vis. It is important that whatever is created as a base can stand alone and tell a story.

Now add a new story to it. Which companies could afford to pay $100K / year more. Who would dominate?

Dominator By Salary

You could add another view. In this case a comparison by percentage of revenue

Percent Revenue

Explore other base stories that lead themselves to rapid production.

What are good channels to find good candidates / how to successfully apply to a visualization job?
What are good ways to test if a candidate is good at d3 or visualization / how to ace your visualization interview?

Lots of us have used D3 and React together, or want to. But, how? What does a good implementation of this look like? What are the pitfalls?

Possible outcome: 5 best practices for using D3 & React together.

How to design and code knowing that more and more browsing is happening on phones and tablets. Topics could include:

• touchscreen interfaces (tooltips and hovering)
• tapping and swiping
• clicking and dragging
• responsive SVG

beyond the chart or the standalone visualization.
How do we create, design and implement data vis centric tools?
How do we evaluate their usefulness or success ?

I am interested in new ways of visualizing source code and execution.

• See scope and context of a particular line of code
• Handle navigating large namespaces
• See "the world" frame by frame https://github.com/NeilFraser/JS-Interpreter

Let's discuss the multiple strategies to use D3 with WebGL, multi-target renderers, headless environment, WebWorker, etc.

Either "Visualizing Music" or "Sonifying Data" or both.

How do games use data visualizations to accomplish different goals, such as:

• Communicating gameplay mechanics
• Visualizing skill vs. challenge matches
• Aiding in decision making processes
• How data visualization can be used to guide attention to enhance (or detract from) gameplay experiences

It's easy to imagine how a given dataset can be visualized, but it takes time to implement each possible way. Here's a Google Doc: Dataset vs. Visualization Matrix. This crosses datasets with visualization techniques. For each cell, if there is an X with a link, then it links to a block on bl.ocks.org that implements that visualization of that dataset. If there is an X that is not a link, then it should be possible to visualize the dataset with that visualization technique, but it just hasn't been done yet.

The proposed session is that we break into teams of two or three and collectively fill in all Xs with links to fresh blocks that we develop on the spot. The end product will be a nice collection of many different visualizations for each dataset listed. During the session, to claim that you are working on a given cell of the table, put your github user name there. This will ensure that there is no duplicated work, and that as many cells of the table get filled in as possible.

Take a popular website, Facebook, Amazon, Linkedin or your favorite or least. Deconstruct the front page. Re-think it in terms of visual elements. For example, would the opening page of Linkedin be more effective with a network layout of your connections. Would Facebook be more effective and engaging with a visual drill down showing your world in one view.

Possible Format: Pick a site. Just focusing on the front page identify 3 things that you would want to highlight or make more graphically accessible to the user. Then walk through a mock-up.

Grab a headline from today news with interesting data and create a vis for it. Using everything that D3 has to offer craft a visualization over the course of this session. This is an exercise of how fast you can create something compelling. One approach could be to have the group point out possible visualization targets with suggestions on what and how to visualize. A designated moderator would then organize these ideas for a designated ace D3 coder to implement. Hopefully by the end of the session you have something cool.

The next version of D3 v4 is all split into tiny modules. It could be used to build custom versions of D3, to share more granular modules, to have a cleaner separation between all components, etc. The future is modular, how can we be prepared?

I have some interest in doing this full time. If we have participants who already do, I'd like to know what skills, knowledge base, and achievements they feel are indispensable for the job.

Using a simple python script to generate d3 visualizations and keep code organized and modular.

Why python? Because I do most of my data gathering, processing, and analysis in python. Inspired by python frameworks like Flask and Django, various templates can be used depending on destination of the visualization. For example, is the project a stand-alone data-art installation, or is it a gist that will be imported into a wordpress site?

Whoever wants to gets 2 minutes to show off some of their bl.ocks. Modeled after the wonderful "show and tell" portion at the end of each Bay Area D3 meetup. Add links here for quick access!

So many people use d3 to create awesome websites but you can also use d3 to manipulate websites that are not yours and do some really wacky things. I'd also like to share a small d3 plugin I've created called d3fx to make defacing websites fun and easy!

We now have bl.ocks.org, bl.ocksplorer.org and blockbuilder.org all helping the d3 example ecosystem.
In this discussion I'd be interested in discussing how to make these tools better, and what other tools we could use/make that might help us.
We've got new insight into the blocks being created by the community that has been explored via color, time and api. Perhaps we could talk about further understanding the examples that the community is creating and how we can facilitate more of that.

Power BI just enabled custom visualizations. We're a SaaS service for business users to bring together data from any data source created by Microsoft. Wanted to share what we're doing, listen to your thoughts, and I hope spark some inspiration.

You can easily plug in your D3 visualizations into Power BI's data stack that handles all the data queries for you. You write you visual once and then you can use it for any data in Power BI, including data from SaaS services like Salesforce, enterprise sources like SQL or Oracle, cloud big data sources like Spark, or real-time application data. It enables you to easily have multiple visuals on a single page, cross highlight, filter, brush and slice. You can event send your custom visual to others so they can use your visual in new and novel ways, without needing to know how to code.

We've made a huge bet on D3 and are eager to both share what’s possible today and to see how we can plug back into the community. We're pretty serious about our OSS projects - we open sourced our visualization stack under MIT License. It's a general purpose visualization library you can use to build re-usable visualizations.

I'd like to share with you what Power BI is, how easy it is to create visuals for Power BI, and walk you through our community driven gallery of custom visuals for Power BI. I also wanted to take time to hear your thoughts and see what's important to the broader D3 community.

I'm super excited to join unconf and meet you all!

Check out these links, I hope they spark your imagination:
Custom visuals in action (video)
Community visuals gallery
Microsoft/PowerBI-visuals playground
Microsoft/PowerBI-visuals (MIT license)
Power BI

D3 gives us the building blocks to build interactive visualizations from the ground up with ultimate flexibility, but are there ways to effectively encapsulate reusable chunks of code that implement common visualization patterns? For example, how many times have you coded the margin convention?

This proposed session will discuss lessons learned and broader goals of the Chiasm Project. This will include a discussion of how functional reactive programming can be leveraged to build composable abstractions for three levels of granularity:

• reactive mixins (D3 margin convention, dynamic resize, scales, axes, brushing)
• visualization components (line chart, scatter plot, bar chart, pie chart, choropleth map)
• interactive visualizations with linked views (linked line chart and bar chart, focus + context plots)

Microscopes let us see tiny things, and telescopes let us see massive things, but what lets us see what is going on in the world at the level of humanity? The broader goal of the Chiasm project, and perhaps of many others within data visualization communities, is to build a machine similar to a microscope or telescope that we can "point" at any domain for which there is data available and see pictures that no one has seen before, giving us new insights into the world in which we live.

One example of this is modifying this visualization of migrant deaths to visualize the recent Paris attacks .

Tamara Munzner has done an amazing job in developing her book Visualization Analysis and Design, which includes a fairly comprehensive mental model for how data can map onto visualizations in terms of visual marks and channels, and what kinds of interactions are possible to augment the static visualizations.

The proposal for this session is to collaboratively assemble a list of blocks from bl.ocks.org that demonstrate each of the visualization idioms and interaction techniques outlined in the book Visualization Analysis and Design. We can use this set of slides outlining the book as our guide.

D3 has 40,000 stars but we don't really have a good sense of the people using it. Let's design the ultimate survey and do a d3.census() that provides a dataset of skill, job types, fields, process, practice and whatever else we can measure about the community. And then we'll have a great open dataset that anyone can use to create awesome data visualization.

Speed is critical in data viz for people to have a good experience and explore the data. But, how do you do it?

Share tips & tricks for making D3 fast!

D3 has a lot of quirks, and often time not as straight forward as other lib. I'm sure we have all encountered some issues and had to find hacks or workarounds.

It would be nice go come together and present the our problems to see if someone in the group has the solutions.

Collaborate on sketching (on paper or in D3) a custom visualization. Bring a dataset or fork a prepared one. Help each other brainstorm and create.

Techniques for tracking the user in a visualization. See what they click on. Is it the best path? Does it make sense to have a best path? Is the vis explanatory or exploratory? Will that determine how you evaluate the effectiveness?

Possible Outcomes:

• Description of a framework that logs all interactions
• How do we measure effectiveness?
• Best practices

How should data visualization and virtual reality work together? What are the opportunities there?