Data Visualization

Hello World: I’ve Graduated

June 15, 2015 by michellechandra in Data Visualization, Photography

Two years ago, I packed my bags and left California for New York. Graduate school ended up being a tough haul, especially attending a challenging technical school when I had very little background in my new subject.

I hesitate to recommend such an endeavor to, well, anyone.

The good news is, I made it! Just a few weeks ago, I published my thesis and graduated to the words of Robert De Niro.

The even better news is that my thesis has resonated with people around the world. “All Our Yesterdays” has received great press. It’s been wonderful to hear from people from all over who have enjoyed the project. It was especially exciting this past weekend to see a writeup from the Daily Mail, which made my parents (who met in England and to this day watch the BBC) super proud.

Here’s a round up of press the project has received in the past few weeks:
Daily Mail – UK, From Dawn to Dusk
Engadget, Mapping project catalogs Instagram sunrises from around the world
Gizmodo, Travel the World Through Sunsets and Sunrises on this Map
PetaPixel, A Visualization of Sunrise and Sunset Photos Being Snapped Around the World
PSFK, See the World’s (Instagram) Sunrises and Sunsets as they Happen
The Creator’s Project, Instagram Viz Project Lets You Catch the Sunrise All Over the World
Design Taxi, You Can Now Watch The Sun Rise And Set Around The World Via Instagram

It’s strange to be let loose back on the world. That is to say, I think it’s time to start new projects!

My San Francisco Photodata

March 29, 2015 by michellechandra in Data Visualization, Photography

I moved to New York City in July of 2013 from San Francisco where I had lived on and off since 2007. In the six months leading up to my departure, I worked two temp gigs in downtown San Francisco and commuted from where I lived in the Outer Richmond. Just prior to these jobs, I caved and bought a smart phone.

I spent my last 6 months in SF photographing the city as a way to alleviate the boredom of my temp gigs and explore. These choropleth maps are abstracted to only show the neighborhoods and blocks where I photographed in SF from January 2013 – July 2013. The maps reveal the neighborhoods I spent most of my time and the locations I found visually interesting.

I ended up traveling a good deal before moving to New York City as I realized once I was in graduate school, I would have less opportunity to travel. This choropleth map shows the photos I took by state in the same timeframe.

I created the maps in CartoDB using geospatial aggregation to group the photos by administrative regions. I was hoping to also include the SF bus routes I took frequently, but I was unable to find a dataset that I could bring into CartoDB.

Generative Star Trails Animation

November 15, 2014 by michellechandra in Data Visualization, Music

For my final for Code of Music, I have decided to change projects and create a generative version of my Star Trails Music Visualization.

While I was living in San Francisco, I frequented a local grassroots music and arts space (Red Poppy Art House) and one day, one of the artists tied to the space, Meklit Hadero, played a few songs on her guitar that used recordings of stars by Nasa’s Kepler Mission for the sound. The star sound recordings are provided freely by Nasa, and I am intrigued by the idea of using them for the generative melody for my Star Trails animation. The star sounds are droney and strange and I think it would be interesting to filter and distort them and create a beautiful and eerie soundscape. Not sure yet if the ocean landscape should also have its own sound tied somehow to data/sonification as well (perhaps I will distort ocean recordings in keeping with the theme!)

I am also interested in recreating the Star Trails animation in 3D and using one of the example three.js particle wave sketches for the ocean. I like the idea of allowing the user an ability to navigate and explore the animation in 3D, or possibly having the generated music change the camera position and the lights. That said, I definitely need to put aside some time this week to dig into the tone.js examples and API in order to understand the library.

Some interesting data sonification projects:

Listen to Wikipedia
Financial Crisis Sonified
London Sound Survey Map
Sonic Water

Programming Color – Hue, Saturation and Brightness

October 26, 2014 by michellechandra in Data Visualization, Music

Successfully set up my Star Trails music visualization to use Hue, Saturation and Brightness as the color parameters.

I am mapping the hue of the stars to a small range (shades of blues) and to vary with changes in frequency values (low, mid and high notes). Color saturation is lower in the beginning so as the “stars” are painted across the sky, the color builds over the duration of the sketch. The saturation and brightness of the colors is also mapped to changes in the volume (when the song is louder, the stars that are drawn are brighter and more vivid in color).

The diameter of the stars also changes according to the frequency spectrum. I am also mapping the stars’ “speed” to the frequency spectrum, which means that rather than having the stars continuously draw a line, there are larger jumps and gaps in the line and moments where the stars are painted in a previous position. The result is non-continuous lines and a more spontaneous and organic look to the piece.

The initial idea of the sketch was to have the stars draw circles which would be completed at the end of the song. The result was very crowded and too perfect. Allowing the music to influence the way the stars are painted across the sky results in more organic variation and results in a musical interpretation of star trail photography.

Sticking to a monochrome color scheme (shades of blue) seems fitting given the relaxing and meditative nature of the sketch. This also creates a harmonious look to the piece, with ocean and sky reading as being part of one sketch. The color of the ocean lines also varies according to changes in the volume, while the range of possible hues is in the blue/purple family. It’s also fun to let the ocean lines be mapped to greater values and more visibly see the resulting change of volume, although it doesn’t work with the meditative quality of the sketch.

Feeling very close to a finished look and feel! Am interested in trying to map the color variables to more than one visual element (for instance, alpha values mapped to both the y and x axis) and continuing to explore what the code is capable of doing.

Star Trails Music Visualization – Color Theory

September 29, 2014 by michellechandra in Data Visualization, Music

Last step before finishing my Star Trails music visualization is getting the colors right. First step in experimenting has been moving away from the RGB color space and thinking in terms of Hue, Saturation and Brightness.

Low frequency values mapped to center, higher values mapped to outer edge

The challenge is in understanding how the music data is affecting the color mapping. Right now, I have only been able to create a monochrome looking sky, when I would rather have more variation in colors. Part of the reason for the monochrome look at present is that the frequency values are mapped to the x axis so that low frequency values are nested in the center and get higher as move outward. Each star represents a different frequency value (low, mid, high). It might be better though to randomly draw the different values in order to result in a more varying sky. It is also possible that because the song being visualized does not change so much, there is less visual variation.

I need to get in the habit of console logging my sketch to make sense of what is going on. Once the colors are closer to where I want them to be for the sky, I will map colors for the landscape lines as well. Right now, the landscape and sky read as two separate visuals rather than being part of the same sketch.

This project has been an ongoing enormous endeavor but thankfully, the visual look of its has vastly improved over the past week now that I am experimenting with hue, saturation, and brightness!

Star Trails Music Visualization

September 21, 2014 by michellechandra in Data Visualization, Music

I am currently developing Star Trails, a music visualization that takes inspiration from long exposure night time photography of stars, whose light trails record the rotation of the earth. Star Trail’s animated nighttime landscape is an abstract representation of music and a celebration of pattern and repetition.

An example of star trail photography

I started this project over the summer, and tapped fellow ITP-er Jason Sigal to help as my understanding of music and sound is limited. The project is a collaborative one with both of us contributing with the coding while I am taking the lead on visuals and concept.

There have been many challenges with this project, challenges perhaps universal to music visualizing in general: → Read more

Goodbye Summer, Hello ITP!

August 31, 2014 by michellechandra in Data Visualization, JavaScript

Looking forward to being back in the incredibly inspiring environment that is ITP – NYU this week!

Last week, I attended a Maptime NYC meetup where Adam Pearce (developer at Bloomberg) walked me through animating with D3.js, and in consequence, improved my understanding of D3 100 million times over. It’s been a slog teaching myself JavaScript and D3 this summer solely through readings, altering code and the occasional Meetup. Having someone explain concepts while working on code was incredibly helpful, and the D3 documentation is making much more sense now.

Here is a US map of my travels created in July using D3:

And a world map with an arc animation effect to connect locations created just last week at Maptime NYC:

There are TONS of resources for learning how to make maps including Mike Bostock’s incredible tutorials, documentation and examples (pretty much everything you can think of to make with D3.js, Mike has made, and nicely coded up for you), meetups including Maptime and GeoNYC, and Scott Murray’s Interactive Data Visualization for the Web.

Looking forward to hopefully finishing these “summer” projects soon, and starting new ones this upcoming semester!

Powered by WordPress - Built by ThemeShift