April 18, 2012
April 15, 2012
We have our first non-interactive visualizations produced in Proclipsing. The different drawing were made my changing properties, everything is very easy to adjust so these are some of the examples that we could show in our main visualization.
Next we need to create a legend, add interactivity by displaying more information on each planet with our "info box" and allow the user to make changes to what is encoded on the main visualization.
This first visualization shows how color is adjusted. Confirmed planets appear RED, being followed on GREEN, candidates GRAY. All planets are mapped according to their coordinates.
This visualization uses the same coordinate system but the colors are based on the color of the star the planet orbits. We will be refining the color encoding, this is just a test.
It is also easy what the axis represents. Here we chose to map how the planets gravity compares against its distance to the host star.
April 12, 2012
After a series of discussions the framework for the visualization is complete. We decided that to make the visualization easier to develop we would take an object oriented approach to the visualization. However the Processing editor was not a great way to construct a multi-object visualization.
Pankaj found a tool called Proclipsing that would allow us to use Eclipse to write Processing applications. Using an IDE would allow us to better manage the code that was being created.
The project is broken into 5 packages each of which contain classes and enums that are used to build the visualization:
Keplerplanets: Sets up and draws the visualization
Keplerplanets.datareader: Reads in all the data to our visualization
Keplerplanets.graph: Creates objects for each planet and draws them on a graph.
Keplerplanets.settings: Stores all the settings for the visualization
April 5, 2012
Based upon Milady’s research on the choice of coordinates, I decided to plot the Declination (dec) vs. Right Ascension (ra) -- which is roughly equivalent to showing a viewer the view from earth and is a one-to-one mapping to point in the space. (Most amateur astronomers use equatorial coordinates to set their telescopes -- Right Ascension is roughly equivalent to a celestial longitude, and Declination like a celestial latitude)
Here’s the visualization in Tableau:
As you can see -- the Kepler measurement array is clearly apparent in the figure. I chose the following color scheme:
Gray -- kepoi_type identified as CANDIDATE.
Blue -- Identified as a candidate that is being followed up in detail.
Red -- Identified as CONFIRMED planets.
For the encoding of size I chose plotting the log of gravitational constant (which comes from data) -- this gives us as size an approximate idea of the gravity of the planet. (Earth’s Gravity = 3 here because this is log of g in cm/s^2, which for earth is approx. 1000 cm/s^2)
This visualization lines up perfectly with the design of the Kepler focal plane assembly:
The visualization allows you to see how many planets there are in each of the 25 individually mounted modules.
April 5, 2012
After a hiatus the group met and we ran through remaining tasks. Milady had finished collecting data on the project and was able to give explanations on the data fields collected. Pankaj and Dave brought the data into Tableau to make the following mockups:
What I have done is plotted Log of (Planet Stellar distance in AU) versus Log of (Planet Radius in terms of earth) --- why?
This way -- Earth in both axis is (0,0) because 1 AU is log(1) = 0 and 1 Earth radius is 0 again when taking log.
This also shows us nicely most points without dense clustering. Interesting patterns emerge by the way -- the most obvious thing that jumps out is that most discovered planets are closer than the earth sun distance! (Visualization score! Yes!! - discovered something that is not obvious from table!)
Also I was thinking about the code plan - It looks like we will need at least the following objects:
- Planet -- this will contain the attributes and draw itself.
- Graph -- this is the graph paper or grid.
- Info box -- this is the box that changes when we mouse over the planets.
After seeing Pankaj’s visualization (which is awesome) I wanted to see if there is a way to show the data in a more linear fashion.
We do not have the date information on when the planets were discovered but after reading up on the data the Kepler ID for each planet is sequential. So I plotted the planet’s distance to its start against its ID. I then encoded the dot size as the mass of the planet and used a color encoding as an indicator of temperature.
I liked the overall effect, it made it easier to see the planets and there was less clustering. We may want to consider the ability to encode different values in the main visualization similar to the homework we had with the NBA players.
I have been focused on data elements to ensure the team has an understanding of data attributes: meaning and use of terms. I have been utilizing the data dictionary below and other web pages to clarify items and answer questions for the team:
- min/max values,
- stellar coordinates (latitude/longitude) different methods to ensure using the correct one.
- color: a substantially complex topic. Color is used in many ways:
- to represent temperature
- to represent mass and temperature
- to represent atmospheric extinction
- to represent a combination of the above.
- color identifiers or classifications
- Stellar Spectral Type is the classification most used to refer to star colors and other related characteristics such as temperature
- g, u, r, i and other scales: Magnitude scales (J-K, gr, gK), used to speak to brightness and distance in light years. Letters refer (in most instances) to filters used to observe stars (green, red, etc).
- discovery date (Exoplanets archive does not contain discovery date for the 2,321 records in the Kepler Planetary candidates file).
- identifiers: kepler id versus kepler name versus kepoi type
- kepler id: sequential number
- kepler name: name for confirmed planets
- kepoi type: classification