Historical

#30DayChartChallenge Day 15

Today’s theme is Historical and I opted for a simple interpretation of showing the changing monthly temperature distribution over a year. Data is taken from a Python version going through what I consider the best time series textbook, Forecasting: Principles and Practise.

import polars as pl
from lets_plot import *
LetsPlot.setup_html()

bikes = pl.read_csv('https://raw.githubusercontent.com/nshahpazov/fpp-in-python/master/data/bikes/day.csv')

daily_temp = (bikes
    .filter(pl.col("yr") == 1) # 2012
    .with_columns(temperature = pl.col("temp")*41)
    .select("temperature","mnth")
)

monthly_averages = daily_temp.group_by('mnth').agg(pl.col('temperature').mean().alias('mean_temp'))

df = daily_temp.join(monthly_averages, on='mnth', how='left')

(ggplot(df, aes("temperature","mnth"))
    + geom_area_ridges(aes(group="mnth",fill="mean_temp"))
    + scale_y_continuous(
        breaks=[1,2,3,4,5,6,7,8,9,10,11,12], 
        labels=['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec']
    )
    + scale_fill_viridis(name="Mean Temperature", option="turbo")
    + ggsize(width=600, height=620)
    + labs(
        x = "Temperature (°C)",
        y = "Month",
        caption = '#30DayChartChallenge #Day15 Historical\nData: https://github.com/nshahpazov/fpp-in-python/tree/master/data/bikes\nMade by: www.ddanieltan.com'
    )  
    + theme(
        legend_position='top',
        plot_caption=element_text(size=12, color='grey'),
        plot_title=element_text(size=20,),
    )
)

TIL

1. scale_y_continuous is the method that allows one to have fine grain control over the range of axis values and labels. I use it here to force my y axis to stay within 1-12 and to rename labels to more recognisable months.

2. If the chart is direct enough, one can omit the title without losing any information! I realised this right before I published, and feel that removing the title takes nothing away from the message and adds more whitespace.

3. Learnt what format the df should be when using geom_area_ridges() by consulting this example page.