pyOpenSci Contributor Metrics

import warnings
from pathlib import Path

import altair as alt
import pandas as pd

# This is a local module that stores the plot theme
from pyosmetrics.plot_theme import load_poppins_font, register_and_enable_poppins_theme

pd.options.mode.chained_assignment = None
pd.options.future.infer_string = True

# Suppress all warnings
warnings.filterwarnings("ignore")

# Load the & register Poppins theme
load_poppins_font()
register_and_enable_poppins_theme()

# Get the current notebook directory
current_dir = Path.cwd()
parent_dir = current_dir.parent.parent
data_dir = parent_dir / "_data"
all_files = list(data_dir.glob("20*"))
dfs = [pd.read_csv(file, parse_dates=["item_opened_by"]) for file in all_files]
df = pd.concat(dfs, ignore_index=True)

df["quarter"] = df["item_opened_by"].dt.to_period("Q")

# Group by quarter and contrib_type (staff vs non-staff)
df_quarters = df.groupby(["quarter", "contrib_type"]).size().reset_index(name="count")
df_quarters["quarter"] = df_quarters["quarter"].astype(str)

# Create a new column for filtered labels
df_quarters["quarter_label"] = df_quarters["quarter"]
df_quarters["year"] = df_quarters["quarter"].str[:4]
df_quarters["qtr"] = df_quarters["quarter"].str[-2:]
df_quarters.reset_index(inplace=True)

# Function to insert newline before the "Q" character
def insert_newline(s):
    return s.replace('Q', ' Q')

# Apply the function to the 'quarter' column
df_quarters['quarter_label'] = df_quarters['quarter_label'].apply(insert_newline)

# Create the Altair chart
chart = (
    alt.Chart(df_quarters)
    .mark_bar()
    .encode(
        alt.X('quarter_label:O')
            .title('Quarter')
            .axis(labelAngle=45),
        alt.Y('count:Q')
            .stack(True),
        alt.Color('contrib_type:N')
            .title("Contributor Type")
            .scale(range=["#81c0aa", "#735fab"]),
        alt.Tooltip(["quarter_label", "contrib_type", "count"])
    )
    .properties(title="Staff vs. Volunteer Contributor by Quarter")
    .configure_legend(
        orient='top',
        titleAnchor='middle',
        direction='horizontal',
        padding=10,
    )
    .interactive()
)

chart