Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Big Data Week: Spatial Data, Visualization, and CartoDB

Big Data Week: Spatial Data, Visualization, and CartoDB

This talk expands upon my previous demo of spatial tools and visualization with CartoDB. Topics covered include design considerations, scientific visualization vs communication, relevance of big data, and a step-by-step walkthrough on basic map creation in CartoDB.

Presented during Big Data Week (http://bigdataweek.com/) at Penn State.

Joshua Stevens

April 24, 2013

More Decks by Joshua Stevens

Other Decks in Science


  1. Table of contents: 1.............Considerations 2.............Big Data 3.............Art, Science, and The

    Scary Places in the Middle 4.............Why CartoDB? 5............Making a Map with CartoDB 6............Q & A
  2. Spatial is special. “Everything is related to everything else, but

    near things are more related than distant things.” - Waldo Tobler (1970) Preface:
  3. Geographic visualization should be forward thinking. Step 1: Look ahead.

    What is your goal? How will you accomplish it?
  4. Big Data are more than just ‘big.’ “Big data is

    more than simply a matter of size; it is an opportunity to find insights in new and emerging types of data and content...” - IBM Step 2: Consider your data.
  5. Step 3: Decide if you’re looking for answers or communicating

    results. (or both) - Andy Woodruff, “Apart from being dead, Art and Science are strong in web cartography.”http://andywoodruff.com/blog/apart-from-being-dead-art-and-science- are-strong-in-web-cartography/ “Web cartography is not about maps; it’s about hacks for moving data around.” Graphic from MacEachren, A.M. (1994) “Visualization in modern cartography: Setting the agenda.” Redrawn by Roth (2011).
  6. So how does this relate to CartoDB? To find out,

    let’s look at what other tools do.
  7. http://at-cam.blogspot.com ArcGIS http://blog.ushahidi.com ✓ De facto standard GIS ✓ 10.1

    on campus labs ✓ 1-year licenses for students ✓ Raster and vector analysis ✓ Hundreds of tools ✓ Model builder (see left) ✓ Python
  8. Quantum GIS (QGIS for short) ✓ Windows, OS X, Linux

    ✓ Open Source ✓ GDAL ✓ PostgreSQL/PostGIS, OSM ✓ Raster and vector analysis ✓ Hundreds of tools ✓ Python ✓ Increasingly mainstream
  9. PostGIS Spatially-enabled PostgreSQL ✓ Windows, OS X, Linux ✓ Open

    Source (GPL) ✓ Very powerful. Very fast. ✓ Often used in conjunction with other tools No direct GUI
  10. R With Spatial Packages ✓ A stats favorite ✓ Excels

    at point patterns, trends, and interpolation ✓ So-so visualization ✓ Recommend packages: ggplot2, maps, splancs, spatstat, mapproj R Studio w/ ggplot2 + maps http://www.statisfaction.wordpress.com
  11. TileMill ✓ A MapBox project ✓ CartoCSS for styling ✓

    Smart selectors + compositing ✓ Shapefiles, PostGIS, SQLite, GeoJSON, CSV... ✓ Export to web, mobile, tiles, PDF, SVG, PNG... ✓ Publish to MapBox (free and $) ✓ Open Source
  12. MapBox ✓ Cloud-based map design ✓ Add public layers, base

    maps ✓ Load data from TileMill ✓ Embed or share ✓ Free account = 3,000 views per month ✓ MapBox.js library
  13. D3 ✓ JavaScript for SVG styled with CSS ✓ Huge

    (and growing!) set of examples ✓ Impressive geographic capability ✓ Steep learning curve ✓ Maps are only one facet of D3 Data-Driven Documents
  14. Where does CartoDB fit in all of this? DB-driven (PostgreSQL,

    PostGIS) Scriptable tools and extensions Support for SHP, CSV, GeoJSON Tie into R for analysis Cloud-based online editor Pair with JavaScript libraries (D3, MapBox.js, etc) Base map from TileMill, MapBox SVG rendering Animation Interaction Communication Analysis Cloud-ready You guessed it!
  15. Let’s make a map! 3 Any plans for academics? “Yes

    we have. Contact us for getting more information. We are quite friends of academics so, you will get a lot of benefits.” - CartoDB
  16. Let’s make a map! Points, lines and polygons = geometric

    primitives essential to GIS + cartography
  17. Let’s make a map! 6 Data type can be String,

    Number (float), Date, or Boolean
  18. Let’s make a map! 1 2 3 CartoCSS Breakdown 1

    2 3 + + +... = multilayer symbol for heat map Smart selectors enable variable-driven design