NF26 - METAR project

Access to pdf report and presentation.

• Keyspace name : finland_weather_metar
• Columns family name for question 1 : spatial
• Columns family name for questions 2 and 3 : temporal

Exercise 1 : stats for a given place

Launch in ipython (not pyspark) :

$ cd nf26-metar
$ ipython
In [1]: %run stats_of_station.py

Examples of available functions :

Once you have run the script, we will show you the available fonctions. You can copy-paste one in ipython and change the parameters.

List of useful indicators :

temperature_fahrenheit, dew_point_temperature, relative_humidity, wind_speed, one_hour_precipitation, ...

---- Stats for a station in range of years

$ requete_interval("EFKI", indicateur="dew_point_temperature", from_year=2006, to_year=2012)

---- Stats for a station in a specific year

$ requete_in_year("EFMA", indicateur="temperature_fahrenheit", year=2010)

---- Stats for a (lon, lat) in range of years

$ from stats_of_station import *
$ requete_lon_lat_interval(28.5, 61.5, "dew_point_temperature", from_year=2006, to_year=2012)

---- Stats for a (lon, lat) in a specific year

$ requete_lon_lat_in_year(28.5, 61.5, "temperature_fahrenheit", year=2011)

Graphs created will be all in folder results/stats

Exercise 2 : map for a given instant

Launch in ipython :

$ ipython
In [1]: %run map_by_indicator_and_time indicator yyyy-mm-dd hh

An instant is considered as a date and an hour. Find the map in results/map_by_indicator. See previous section for a list of interesting indicators.

Exercise 3 : stations clustering for a given period of time

Launch in pyspark :

$ source /pyspark.env 
$ pyspark
In [1]: %run cluster_by_period yyyy yyyy

A period is considered as a range of years. The first argument is the start year, the second one is the end year. You should better specify an end year greater than 2009 as we only have data for one station until 2010. Find the map in results/cluster_by_period.