Ducky's Info Vis Milestones

AIs

• try caching with the mapki technique
• look into changing the color table in PNG files with gd
• make test4.html dynamicly generated -- use window.location.search to get the query string
• use arg0 to determine where the min and max cutoffs are; link to all those

Current worries

• speed, speed, speed
• What are storage needs?
  • in Palo Alto, .005 degrees long for 500 pixels at max zoom in => 100,000 pixels / degree
  • in Palo Alto, .002 degrees lat for 300 pixels at max zoom in => 150,000 pixels / degree
  • US is about (-66 to -124) 58 degrees wide => 12,000 images wide (at 5K apiece)
  • US is about (32 to 48) 16 degrees tall => 8,000 images tall (at 5K apiece)
  • so at max zoom level, 96M images to cover the whole US * 5K => 480G
  • roughly powers of two as zoom out; 15 zoom levels before whole US fits (after that, it's just silly?)
  • so roughly 2*480G worth of space needed for whole US, or about 1000G
  • if it takes about 5 sec to render a little one =>

• What are units of areas?
• What if different browsers don't work? -- tant pis for now
• caching
• How do I deal with the huge dynamic range?
  • Offer several:
    • Which?
    • How does user switch between them?
• Do I allow people to change range on the fly?
• How do I capture clicks through the overlay (the map/satellite lens uh-oh does things differently).
• Speed speed speed
• LD_LIBRARY_PATH / static libraries nonsense

Ducky's Info Vis Tasks

• Explore different color bars
• Thoroughly understand and document how census bureau data fits together

• Develop a simple test suite for PNG manipulation
• Write code to create PNG from one known polygon extracted from shapefile; overlay on Google Maps
• Write code to create PNG from several known polygons extracted from shapefile; overlay on Google Maps
• Write code to create PNG from several known polygons, shaded with data from population file; overlay on Google Maps

• Given zoom level and map lat/long boundaries, figure out how to tell which polygons are visible
  • test better
• Write code to create PNG from lat/long boundaries at one zoom level
• Write code to retrieve a polygon based on lat/long.
• Port code to Dreamhost
• Plug code into CGI
• Figure out how to redraw with panning
• Write code to deal with panning via controls
• Figure out how to drag the overlay (ulp, what if not possible?)
• Figure out how to aggregate data from multiple polygons -- maybe not needed?
• Write code to aggregate data
• Modify code to create PNG from lat/long/zoom data
• Overlay images dynamically on Google Maps

• (aggressive) Add sliders to allow changing the parameters of the image (e.g transparency, hue)
• (aggressive) Add radio buttons to change data sets
• (aggressive) Allow user to choose from several different years of data
• (extremely aggressive) allow semi arbitrary data to be mapped
  • design with XML format
  • design interactive
  • find information about Mike Beckerle's data description language

Proofs of concept:

• simple walking duck
• two overlays
• overlay transparent Google Maps over Google satellite images

Promised milestones

• 4 November: Proposal finished
• 13 November: One PNG file generated with one polygon extracted from dataset, requiring:
  • extracting polygons with the Shapefile C library
  • extracting the appropriate demographic data from the dataset with the Shapefile C library and the data files
  • associating polygon <->demographic data
  • understanding the gd library well enough to generate a PNG with a transparent polygon
• 20 November: Functional panning, perhaps limited to one US state and perhaps slow, requiring:
  • determining which polygons will be visible given lat/long bounds
  • iterating through visible polygons, drawing each
  • plugging overlays into the Google Maps API
• 4 December: Functional zooming
  • determining which polygons will be visible given lat/long bounds and zoom level
  • aggregating lower-level data (note that I do not think this is trivial)
  • killing small code gremlins
• 11 December: Performance tuning and bug fixes done, probably requiring finishing:
  • pre-processing census bureau data into a form that is easier to extract quickly
  • prefetching of neighboring overlays
  • caching previously-used overlays
  • killing large code gremlins
• 16 December: Status update due, requiring
  • bibliography done
  • writing done
• 19 December: Final due date, requiring
  • learning LaTeX
  • coercing text into journal format
  • killing LaTeX gremlins

• Set ALLOWTOPICCHANGE = DuckySherwood