NOTE, not used by Viltsitgen (rpi1) anymore, kept for reference

Wind map for Sweden

This wind map for Sweden is an exercise using D3 and REST. It is modelled after from Peter Cook's UK Wind Map and loosely inspired from Wind.

In short, there is a python script (getSMHI.py) that traverse the SMHI REST API and download wind speeds and directions from 'SMHI latest day' data. This data is rendered through D3 on a SVG map of Sweden. Min, max and average wind speed is calculated as well as average wind direction (in degrees).

Details as follows.

Python

The Pythons script getSMHI.py is located in the directory py and is using the documentation available on this link http://opendata.smhi.se/apidocs/metobs/ The entry point is hard coded in the python script as is the parameter values for wind speed (4) and wind direction (3). Data is downloaded in JSON, so the 2 JSON structures is combined into one common JSON structure and saved as a file with the date of execution of the script , for example 2015-07-13.js. The file is then copied to wind.js, so this file always contain the latest data. The data in this file is accessible through the variable weather.

The JSON files are stored in the data directory, parallel to the py directory. Files older than 7 days are removed.

Note codes are hardcoded and there is no error checking in the script. Not my proudest script obviously...

Specifics

In my case, the Python script is executed once per day on my Raspberry Pi. I have added the following line in etc/crontab

00 7 * * * pi python /home/pi/app/Swind/py/getSMHI.py

Remember $ sudo chmod g+x -R getSMHI.py and $ sudo chmod a+w /var/log/Swind.log

I changed the ownership of the /var/www/ directory from root to Pi through sudo chown -R pi /var/www/.

If the system time (use date at the Raspberry command prompt) is configured to the wrong timezone, us the following command sudo dpkg-reconfigure tzdata to correct.

To enable git push to GitHub repository, SSH needs to be configured, refer to these GitHub pages:

SVG map of Sweden

This was easily done by following Let's make a map by Mike Bostock, the person behind D3. The shape files and resulting JSON files (the end result is swe.json) is located in the map directory. I modified the instructions by Mike Bostocks as follows.

The first command will convert the shape file to GeoJSON format, the output file being subunits.json and places.json while ne_10m_admin_1_states_provinces.shp and ne_10m_populated_places.shp is the input file.

ogr2ogr -f GeoJSON -where "ADM0_A3 = 'SWE'" subunits.json ne_10m_admin_1_states_provinces.shp
ogr2ogr -f GeoJSON -where "ISO_A2 = 'SE' AND SCALERANK <8" places.json ne_10m_populated_places.shp

Rewrite, not county's... $ topojson -o se.json --properties name=NAME -- subunits.json

This is of course to include Sweden (SWE ISO 3166-1 alpha-3 3 letter code and SE ISO 3166-1 alpha-2 2 letter code) from the shp-file using the -where condition.

Then convert from GeoJSON format to TopoJSON (see Wikipedia article on GeoJSON) format using below, that combines subunits.json and places.json into one output file swe.json

topojson -o swe.json --id-property postal --properties name=NAME -- subunits.json places.json

Here I wanted to include the postal property in the resulting JSON file (swe.json), although it turned out I didn't need it in the end. Originally I wanted to have different colors for each county in Sweden but it turned out nicer to have one color instead.

Animation of Wind speed and direction

ìndex.html includes all logic to generate the wind map. Specifically it includes a D3 mercator project of Sweden

var projection = d3.geo.mercator()
    .center([14.6, 62.1])  // Somewhere in the middle of Sweden, https://sv.wikipedia.org/wiki/Sveriges_geografiska_mittpunkt
    .scale(1000)
    .translate([width / 2, height / 2]);
var path = d3.geo.path()
    .projection(projection);

The function wind_vector takes the current wind speed and wind direction data and scales a line proportionally. It includes other information, such as delay and duration for animation. The lines are projected for the map and stored in the array vectors.

Finally the D3 library is used to render the wind map, refer to the javascript (commented) in index.html for details. In the process the swe.json file is read, the map of Sweden is rendered and then the wind vectors is animated. The average wind speed and direction is drawn in a circle, scaled by the wind max speed. Some text are rendered.

Links