Database Reference
In-Depth Information
...
<parameters applicable-location="point1">
<temperature type="maximum" units="Fahrenheit" time-layout="k-p24h-n8-1">
<name>Daily Maximum Temperature</name>
<value>38</value>
<value>33</value>
<value>41</value>
<value>41</value>
<value>35</value>
<value>32</value>
<value>30</value>
<value>35</value>
</temperature>
<temperature type="minimum" units="Fahrenheit" time-layout="k-p24h-n7-2">
<name>Daily Minimum Temperature</name>
<value>22</value>
<value>28</value>
<value>34</value>
<value>22</value>
<value>24</value>
<value>17</value>
<value>20</value>
</temperature>
</parameters>
</data>
</dwml>
The volunteers at OpenStreetMap have created a somewhat-chaotic but comprehensive set of
geographic information, and you can download everything they've gathered as a single massive
ile. One unique strength is the coverage of areas in the developing world that are absent from
commercial databases, and since it's so easy to change, even US locations are often more up-
to-date with recent changes than more traditional maps. The downside of the system is that it's
designed for navigation, not analysis, so a lot of information about administrative boundaries and
other nonphysical attributes is missing. The
Nominatim project
attempts to organize the data into
a form you can use to look up street addresses, but the lack of good coverage of things like postal
codes limits its usefulness. Reconstructing some structure from the soup of roads and points is
also computationally very taxing, easily taking a couple weeks of computation time on a high-
end machine.
If you're only working in a more limited geographic region, you may want to look at the
Cloud-
made extracts
,
which contain subsets for diferent areas and attributes.
