The State of Open Geo Hacking

From DevSummit
Revision as of 21:30, 20 May 2015 by Vivian (talk | contribs) (1 revision imported)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Facilitated by Evan Henshaw-Plath,

Evan will reprise his ever-popular sessions on geo hacking, offering his thoughts on what's new and cool and then opening the floor up for questions and discussion.

There are two kinds of computer geography folks GIS and Online 3 years ago was the "pushpins on maps era" GIS thought that was stupid b/c it was missing polygons, etc.

TIGER/line = Govt. map data from the census bureau for streets and rivers that's used as a basis for US Street maps

For optimization, tile images are used. Everyblock uses their own tiles

Modest Maps Open source mapping system in flash under bsd license

MSNBC Hurricane tracker built on Modest Maps

Mailing List: GeoWanking


  1. GeoRSS, GeoAtom, GeoJSON - good for a feed of points and also shapes, but not necc complex shapes.
  2. KML - supported by GoogleEarth, OpenLayers, Geo

Can also take an existing feed and extract geographic information from it through something like Yahoo! Pipes.

HealthCity of LA has lots of data that they make available for overlaying various information (i.e. asthma rates vs. income vs. avg commute time to work)

Alternatives to Google Maps

OpenSource Map Stack Drupal article here

More Links

Discussion on how to determine where an address is within a polygon

Geocoders convert addresses into lat/long (often us specific) Google and Yahoo! offer them.,

When storing goespatial data in and RDBMS there is a separate datatype for that data. The best open source database for geospatial data is PostGIS, which is an extension for PostgreSQL


geo mapping

2 fields - geographic information systems (gis) (cartographers), self taught hackers (have taken and restructured the world)

geowankers - mailinglist of geo hackers

from gis to crappy maps like mapblast

australian company that was eaten by google, then launched google maps. not an open platform initially, but people started reverse engineering, using.

pushpins on maps became really easy. all the gis community thought that was stupid, wanted polygons, etc.

gis community

gis - information related to geospatial features - elevation in area as a field, attributes of an area (i.e. population)

portland housing authority - level of income and green space - map layers - see correlation of these two factors

things to figure out - map data, points, lines, polygons. attach all sorts of data to points, lines, polygons

tigerline - used as a baseline for building maps

optimizations that make it more complicated. google stores tiles, loads images as you zoom around.

everyblock - rendering their own -

archaic maps of london - maps of coal mining country that show 1970 view and current view. also

georrs, geojson - rss with additional info -

if you don't have the geo info and have a feed, you can run the feed thru yahoo pipes and have it find geo info and then display on a map.

for the basic stuff, georss is enough.

item in a feed can have a list of points -

KML - defacto standard, can drop it into openlayers. not as supported on the desktop GML - scarier

maps are not things we watch, it is things we create. we can expose data...

LA ngo - obtained lots of data, some open, some they purchased. can overlay all sorts of data on a map.

geocommons mapmaker - can take a snapshot of collection of data and publish it out as a new map.

twitter data was used to do this. clear that it represented a certain


simulation of

dc had an open project - that shows all sorts of data.

geocommons - attempt to collect huge data set

barrier to mapping - developing large enough dataset so you can start - adding overlay of income would give - want to see this overlayed with racist

search for msnbc hurricane maps - stamen design, in sf mission - have their own platform (flash based)

britekite - location twitter - social network update kind of thing that include geo info. probably 40 of them.

search for modest maps - oakland transit data - took pdf and turned into tiles

chicago transit data - available to google but

mapstraction - javascript layer that gives you the same control.

openlayers - open source - initial design is ugly - you do have full css control

national atlas, us govt,

openstreetsmap - attempt to geomap the world - wiki type approach to

beijing - poorly mapped on google, really well mapped on openstreetmaps

styling openstreetmap data - have to do it on the level you render the tiles. mapnik is one tool to do that.

if you want to make a map that does not look like google, changethe style. also, allows you to emphasize info that is not roads.

bike paths and bike fatalities? nyc has crash dat maps, although proprietary with data. real heat maps.

hillmaps in sf also really great.

giving directions - another level of complexities. what direction the street goes in. attribute data is attached to streets. openstreet maps has those attributes. you can redo the routing. google now has car, walk and public transit

mapping people into congressional districts - egocode address, CONTAINS query to find out in which polygon. challenging because of things like

storing geographic data in relational databases. store lat and latitude in a geospatial

best opensource database to store data is postgis

mysql spatial extensions work kinda well. can't talk about the shape, just the rectangle.

for stuff as precise as congressional districts you need to use postgis

biggest trick - 4 sided polygon is 5 points, not 4. it has to go back to the beginning.

fireeagle - built for yahoo - collects info from cellphone

land trust accredation mapping - clustering icons for when many points overlay andrew turner slide presentation

 - - 

onion video thing with slippy map

mountain biking site - ride with gps units, gathering data that will be shareable

geobabble - take gps data and put it into any format you can think of

flicker approach - zipcodes people do not know, visual approach to gathering geo data

geonames georss

reason for a flash implementation is that javascript limit is about 1,000. flash is about 10,000. 100,000 you would use google earth. - lbs- location based services