0. Loading Libraries

import pandas as pd
from dfply import *
from plotnine import *
import ssl

1. Data Loading

## adding below ssl line as it is giving ssl error locally in my machine
ssl._create_default_https_context = ssl._create_unverified_context

groundhogs = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2024/2024-01-30/groundhogs.csv')
predictions = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2024/2024-01-30/predictions.csv')

2. Data Transformation

merged_df = pd.merge(predictions, groundhogs, on="id", how="left")


predictions_groundhog = (merged_df >>
                         select(~X.description, ~X.details, ~X.source, ~X.current_prediction,
                                ~X.image, ~X.slug, ~X.name, ~X.predictions_count, ~X.active) >>
                         filter_by(X.is_groundhog == True))

3. Plotting

3.1. Consensus of Groundhogs YonY

Code for Density Plot of Consensus

plot = (predictions_groundhog >>
        filter_by(X.shadow.notna()) >>
        ggplot(aes(x='year', color='shadow')) +
        geom_density() +
        scale_color_manual(name='Prediction',
                           labels=['Early Spring', 'Long Winter'],
                           values=['#8B4513', '#0000CD']) +
        theme_minimal(base_size=14) +
        theme(legend_position=(0.5, 0.5)) + 
        theme(figure_size=(6,4)) + 
        labs(title='Consensus of Groundhogs YoY'))

print(plot)

3.2. Count of GroundHogs YonY

Code for Line Plot of Groundhogs

plot = (predictions_groundhog >>
        filter_by(X.shadow.notna()) >>
        group_by(X.year) >> 
        summarize(n = n(X.year)) >>
        ggplot(aes(x='year', y='n')) +
        geom_line() + 
        theme_minimal() + 
        labs(x='year', y='number of groundhogs', title='Count of Groundhogs by Year'))

print(plot)

3.3. Consensus of GroundHogs YonY by ShortName

Code for Consensus Plot of Groundhogs by ShortName

plot = (predictions_groundhog >>
        filter_by(X.shadow.notna()) >>
        ggplot(aes(x='year', color='shadow')) +
        geom_histogram() +
        labs(title='Consensus of Groundhogs YoY') +
        facet_wrap('~shortname') +
        scale_color_manual(name='Prediction',
                           labels=['Early Spring', 'Long Winter'],
                           values=['#8B4513', '#0000CD']) + 
        theme_minimal(base_size = 14) + 
        theme(figure_size=(20,12))
                           )

print(plot)

I believe this is a simplified representation of the consensus to gain initial insights.

Considering the number of groundhogs included in the observational study, the majority predicted an early spring.
However, since we currently lack weather data, we are unable to cross-verify which predictions were accurate.
With the exception of Phil, most others forecasted an early spring.