plot_with_plotly.py

#==============================================================================
# Do simple plots through plot.ly, such as line plot, histogram, monthly means
# Function names:
#   timeseries_plot()
#   wind_rose_plot()
#   species_histogram()
#   monthly_box_plots()
#   hourly_box_plots()
#==============================================================================
# Uses modules:
# datetime, numpy, pandas, plot.ly, source_AQ_data, windrose, quick_tools,
# calendar
from datetime import datetime, timedelta
import pandas as pd
import numpy as np
import source_AQ_data
import plotly.plotly as py
import plotly.graph_objs as go
import quick_tools
import calendar
#==============================================================================

def timeseries_plot(species='None',filename = 'ExampleData', average = 'None',
    verified = True):
    """
        Produces a simple line plot of concentration against time.
        Function IN:
            species (REQUIRED, STRING):
                The name of the species you want to plot.
            filename(OPTIONAL, STRING):
                The name of the raw data file, if left it just uses
                example data
            average(OPTIONAL, STRING):
                Choose what type of averaging to use. Choices are None (default),
                8-hour, daily, weekly, monthly
            verfied(OPTIONAL, BOOLEAN):
                Choose whether to plot just verfied data or all data.
                Default = True
    """
    # Get the data for the species required. Also include the filename
    # if the filename is provided - use example data if not.
    # Also returns variable name (this might be changed slightly from user input)
    species_data, variablename = source_AQ_data.select_one_variable(species, filename)

    # If just using verfied data then purge unverified data
    if verified:
        species_data = source_AQ_data.purge_unverified(species_data, variablename)

    # Set the data in a format the plot.ly needs to work
    data = go.Scatter(x = species_data.index, y = species_data[variablename],
        mode = 'markers')
    data = [data]

    # Set the layout for the plot.ly graph
    layout = go.Layout(
        title = 'Concentration of ' + variablename + ' at Edinburgh St leonards',
        xaxis = dict(title = 'Date and Time'),
        yaxis = dict(title = variablename + ' ('+species_data['Unit'][0]+')'))

    # Set the filename & combine data and layout
    filename = '%s_timeseries' % variablename
    fig = go.Figure(data = data, layout = layout)
    py.plot(fig,filename = filename)

    pass

def wind_rose_plot(filename = 'ExampleData'):
    """
        Procuduces a wind rose plot of wind speed and direction.
        Function IN:
            filename (OPTIONAL, STRING):
                The filename of a csv file where this data is kept. If not
                provided then uses the example file.
    """
    # Import the windrose module
    from windrose import windrose

    # Get the wind direction and the wind speed from the file
    wd, wd_name = source_AQ_data.select_one_variable('Modelled Wind Direction',
        filename = filename)
    ws, ws_name = source_AQ_data.select_one_variable('Modelled Wind Speed',
        filename = filename)

    # Combine the two DataFrames
    wind = pd.concat([wd,ws], axis = 1)
    # Drop NaNs
    wind.dropna(inplace = True)

    # Format the windspeeds and directions into windrose format
    dirc_bins = 8
    speed_bins = 4
    windrose_data, speed_bin_names = windrose(wind['Modelled Wind Speed'],
        wind['Modelled Wind Direction'], direction_bin_size = dirc_bins,
        speed_bin_size = speed_bins)


    # Make each circle of the windrose a "trace" for plot.ly
    colours = quick_tools.get_colours_rgb(num_colours = len(speed_bin_names))
    # Need to do it in reverse order so the larger ones don't cover the smaller
    trace_dict = {}
    finished_data = []
    for x,sn in enumerate(reversed(speed_bin_names)):
        trace_dict[sn] = go.Area(
            r = windrose_data[sn],
            t = windrose_data['Direction'],
            name = sn,
            marker = dict(color=colours[x]))

        finished_data.append(trace_dict[sn])

    # Set the layout options
    layout = go.Layout(
        title = 'Wind Speed Distribution at Edinburgh St Leonards',
        font = dict(size = 16),
        legend = dict(font = dict(size = 16)),
        radialaxis = dict(ticksuffix = '%'),
        orientation = 270)

    filename = 'Wind Speed Distribution at Edinburgh St Leonards'
    fig = go.Figure(data = finished_data, layout = layout)
    py.plot(fig, filename = filename)

    pass

def species_histogram(filename='ExampleData', species = 'None', verified = True,
    num_bins = 50):
    """
        Produces a histogram of concentrationof a given species.
        Function IN:
            species (REQUIRED, STRING):
                The name of the species you want to plot.
            filename(OPTIONAL, STRING):
                The name of the raw data file, if left it just uses
                example data
            verfied(OPTIONAL, BOOLEAN):
                Choose whether to plot just verfied data or all data.
                Default = True
            num_bins (OPTIONAL, INTEGER):
                Choose the number of bins for the plot. Default = 50
    """
    # Get the data for the species required. Also include the filename
    # if the filename is provided - use example data if not.
    # Also returns variable name (this might be changed slightly from user input)
    species_data, variablename = source_AQ_data.select_one_variable(species, filename)

    # If just using verfied data then purge unverified data
    if verified:
        species_data = source_AQ_data.purge_unverified(species_data, variablename)

    data = [go.Histogram(x = species_data[variablename],
        nbinsx = num_bins)]
    layout = go.Layout(
        title = 'Histogram of %s concentration at Edinburgh St Leonards' % variablename,
        yaxis = dict(title = 'Frequency'),
        xaxis = dict(title = variablename + ' ' + species_data.Unit[0]),
        showlegend = False)

    fig = go.Figure(data = data, layout = layout)
    filename = 'Histogram of %s at Edinburgh St Leonards' % variablename
    py.plot(fig, filename = filename)

    pass

def monthly_box_plots(filename = 'ExampleData', species='None',
    verified = True):
    """
        Produces box plots of mean, std dev, percentiles of species per month.
        Function IN:
            filename (OPTIONAL, STRING):
                The filepath and name where the csv file is stored. If not path
                is given the the example data is used.
            species (OPTIONAL, STRING):
                The species you wish to plot. If this is left as 'None' then it
                will go to the selecting tool. (Will also do this if the species
                written is not available.)
            verified (OPTIONAL, BOOLEAN):
                Choose whether to use all the data (False) or just the verified
                data (True). Default is True.
    """


    # Get the data for the species required. Also include the filename
    # if the filename is provided - use example data if not.
    # Also returns variable name (this might be changed slightly from user input)
    species_data, variablename = source_AQ_data.select_one_variable(species, filename)

    # If just using verfied data then purge unverified data
    if verified:
        species_data = source_AQ_data.purge_unverified(species_data, variablename)

    # Get a list of months
    month_names = [calendar.month_name[x] for x in range(1,13)]

    # Create dictionary with each month as a key containing all monthly data
    monthly_dict = {}
    for n,month in enumerate(month_names):
        monthly_dict[month] = species_data.loc[species_data.index.month == (n + 1)]

    # Create list to put in the plot.ly traces and combine them
    # to send to plot.ly
    box_data = []
    for x, month in enumerate(month_names):
        box_data.append( go.Box(
            y = monthly_dict[month][variablename].values,
            name = month))

    layout = go.Layout(
        yaxis = dict( title = variablename + ' ' + species_data.Unit[0]),
        showlegend = False,
        title = 'Monthly Averages for '+ variablename +' at Edinburgh St Leonards')

    # Create filename and send to plot.ly
    filename = 'Box Plot of Monthly Average %s' % variablename
    fig = go.Figure(data = box_data, layout = layout)
    py.plot(fig, filename = filename)

    pass

def hourly_box_plots(filename = 'ExampleData', species = 'None', verified = True):
    """
        Produces box plots of mean, std dev, percentiles of species for
        hour of the day.
        Function IN:
            filename (OPTIONAL, STRING):
                The filepath and name where the csv file is stored. If not path
                is given the the example data is used.
            species (OPTIONAL, STRING):
                The species you wish to plot. If this is left as 'None' then it
                will go to the selecting tool. (Will also do this if the species
                written is not available.)
            verified (OPTIONAL, BOOLEAN):
                Choose whether to use all the data (False) or just the verified
                data (True). Default is True.
    """

    # Get the data for the species required. Also include the filename
    # if the filename is provided - use example data if not.
    # Also returns variable name (this might be changed slightly from user input)
    species_data, variablename = source_AQ_data.select_one_variable(species, filename)

    # If just using verfied data then purge unverified data
    if verified:
        species_data = source_AQ_data.purge_unverified(species_data, variablename)

    # Get a list of hours
    hour_names = [x for x in range(24)]

    # Create dictionary with each month as a key containing all monthly data
    hourly_dict = {}
    for n,hour in enumerate(hour_names):
        hourly_dict[hour] = species_data.loc[species_data.index.hour == (n)]

    # Create list to put in the plot.ly traces and combine them
    # to send to plot.ly
    box_data = []
    for x, hour in enumerate(hour_names):
        box_data.append( go.Box(
            y = hourly_dict[hour][variablename].values,
            name = str(hour).zfill(2)))

    layout = go.Layout(
        yaxis = dict( title = variablename + ' ' + species_data.Unit[0]),
        xaxis = dict( title = 'Hour of the Day'),
        showlegend = False,
        title = 'Hourly Averages for '+ variablename +' at Edinburgh St Leonards')

    # Create filename and send to plot.ly
    filename = 'Box Plot of Hourly Average %s' % variablename
    fig = go.Figure(data = box_data, layout = layout)
    py.plot(fig, filename = filename)


    pass


if __name__ == '__main__':
    #data = source_AQ_data.select_one_variable('Ozone')
    wind_rose_plot()
## ============================================================================
## END OF PROGAM
## ============================================================================