Skip to content

Latest commit

 

History

History
165 lines (128 loc) · 4.74 KB

presentation.md

File metadata and controls

165 lines (128 loc) · 4.74 KB

Open Civic Data:

citizens! police! maps!

Maptime HRVA

2015-05-13, Williamsburg, VA

HRVA

goals for today

we will show you

  1. Baltimore case study with 1. riots 1. arrests 1. fatal encounters with police

riots

  • Attribution: TIME magazine
  1. spatial data sources for the US and the world

goals for today

you will

  1. learn how to find and use open spatial data
  2. understand the difference between raster and vector
  3. open GIS layers like .tif ( raster ) and .shp ( vector ) in QGIS

you will

  1. perform attribute joins
  2. eg, join census tables by geo_id
  3. perform spatial joins
  4. eg, counts of arrests by census block group
  5. subset spatial data
  6. eg, find all homicides in the BPD arrest record
  7. reproject vector datasets
  8. eg, between NAD83/UTM 18N and WGS84!

you might choose to . . .

  1. use the American Community Survey (ACS) through censusreporter.org
  2. calculate
  3. heatmaps of incident points
  4. spatial join crime/riots/arrests on census block groups

#heatmap of homicides

if we get to it ...

  1. summarize high resolution land-cover by census blocks
  2. eg, percent tree over by census blocks
  3. this is called "zonal statistics"
  4. cartography
  5. drape layers over DEM products like hillshade

small stuff you'll learn

  1. what is . . . ?
  2. JSON ( hint: open with a text editor )
  3. TSV ( tab separated values - open in QGIS with TAB delimiter! )

Getting started:

download the data

lots of free and open data is available online

like police camera locations

or homicides

OSM (https://www.openstreetmap.org)

osm

OSM (https://www.openstreetmap.org)

osm

Census Reporter http://censusreporter.org

censusreporter

Crowd sourced data curated by @VinylFox and @Ryan_J_Smith

bmore_riots

USDA NAIP Imagery

NAP

Generally data needs to be cleaned or re-organized before it is easily used.

We've done this ahead of time so you can get right to work!

please download the zip file of this workshop at our github repo: github.com/KAPPS-/maptime_hrva

To learn how to bring spatial data into a GIS we will make a simple map with multiple data types:

demomap legend

appendix

BONUS POINTS

summary of example workflow

  1. download NAIP imagery
  2. download NED DEM (digital elevation model)
  3. download censusreporter.org shapefile
  4. download censusreporter.org CSV for another variable
  5. open QGIS, add censusreporter inputs
  6. right click -> properties -> joins
  7. join CSV to shp using census geo_id field
  8. delete Baltimore wide polygon with QGIS edit session
  9. not needed! obscures view!
  10. right click -> save as -> geoJSON
  11. just so we can keep nice field names
  12. right click -> styles -> choose variable and color ramp
  13. overlay points of interest like riot incidents, arrests, homicides, police cameras, etc

CartoDB - make web maps so fast you won't believe it

OpenRefine for the win

  1. text faceting
  2. got an XLS with free text fields gone awry? 1. "Clusters saved my life"
  3. numeric faceting

Why we don't use decennial census

US Census 2000 used short and long forms, but Census 2010 only used the short form. awesome variables like income are now collected using the American Community Survey (ACS) instead

ACS

  • 1 year ( most current, small sample size, only biggest communities )
  • 3 year
  • 5 year ( least current, huge sample size, all areas )

Which do I choose?