Corey Goldberg - April 2007
The Yahoo Maps AJAX API is one of the options available from the Yahoo Maps Web Services Developer API's. Why Yahoo Maps? Because every mashup I see out there uses the Google Maps API. I wanted to try something different, I like the look of the Yahoo maps, and the Yahoo API's are very easy to use and well documented.
We want to create a map of the US and plot our marker image at specified locations. We use the createYahooMarker method to create a new YMarker object for each data point, and add it to our map using the addOverlay method. This is where we define the size of the marker bubble and its location on the map. The API is pretty flexible and offers several options for specifying locations, including city/state, metro, address, zip code, etc:
map.addOverlay(new createYahooMarker('Boston, MA', 13))
(For more info, see the (Version 3 API Reference)
The mashup I describe later needs a pre-processing step that parses and aggregates a lot of data. Python is a scripting language well suited for this. At the end of the script's processing, I just generate the markup (and embedded code) necessary to work with the API.
Static GIF image that gets resized and added as an overlay to be rendered on the map when the page loads:
This is the data file we use in the script above (newline delimited ascii text file).
02116 Dallas, TX Los Angeles, CA Mexico City Havana 03102 Providence, RI Albany, NY Worcester, MA Toronto Edmonton Calgary Atlanta Ottawa Augusta, ME Philadelphia Richmond, VA Seattle
Now that I have shown how to build a simple map with image markers from a data source, we can use this same technique to create a [near] real-time map of user concentration based on IP addresses.
You can also obtain IP addresses of user requests from your server logs. I use a small parsing script to pull out the IP addresses to use as input data.
IP Geolocation data is available from various sources; some free and some commercial. Hostip.info is a community-based project to geolocate IP addresses, making the database freely available. We can use the Hostip API to do our geolocation lookups. We provide an IP address and use the API to translate it into a city/location. Sample code for doing this is shown below.
We start with a list of IP Addresses and convert them to cities/locations by sending a request to the web service API and parsing the response for relevant data.
#!/usr/bin/env python # Corey Goldberg, April 2007 (email@example.com) import urllib import re ip_addresses = ( '184.108.40.206', '220.127.116.11', '18.104.22.168', '22.214.171.124', '126.96.36.199', '188.8.131.52', ) for ip in ip_addresses: response = urllib.urlopen('http://api.hostip.info/get_html.php?ip=%s' % ip).read() m = re.search('City: (.*)', response) if m: print m.group(1)
When we place markers on the map, we specify the size:
map.addOverlay(new createYahooMarker('Boston, MA', 40)) map.addOverlay(new createYahooMarker('Manchester, NH', 20)) map.addOverlay(new createYahooMarker('Dallas, TX', 30))
So the basic approach is to aggregate the data and assign a size value to each marker. We want to render a marker bubble on the city/location with a size that is relative to the aggregate value we assign. To do this, I create an algorithm that chooses a size for each marker based on thresholds I supply.
I have provided an overview of the concepts necessary to create this mashup. Building your own is left as an excersize for the reader.