2 Where do donors live?

How far do Hubbard Hall’s donors live from the hall? Distances are straight-line (haversine) from the venue (25 E Main St, Cambridge NY) to each donor household, bucketed into half-open mile bands. Donors whose only address is a PO box are placed at their ZIP’s centroid and flagged “approximate” (excluded from the tight 1-mile count so it stays honest).

Code

import pandas as pd
import plotly.express as px
from hh.io import read_parquet
from hh.viz import style, BLUE, money, pct

households = read_parquet("processed", "households_summary.parquet")
donors = households[households["is_donor"]]
by_band = (
    donors.groupby("distance_band", observed=True)
    .agg(
        households=("id", "count"),
        dollars=("total_donations", "sum"),
        mean_gift=("total_donations", "mean"),
        median_gift=("total_donations", "median"),
    )
    .reset_index()
)
by_band["share"] = by_band["dollars"] / by_band["dollars"].sum()
by_band["cum_share"] = by_band["dollars"].cumsum() / by_band["dollars"].sum()
by_band.style.format(
    {"dollars": money, "mean_gift": money, "median_gift": money, "share": pct, "cum_share": pct}
)

	distance_band	households	dollars	mean_gift	median_gift	share	cum_share
0	[0, 1)	164	$119,310	$728	$92	8.2%	8.2%
1	[1, 5)	182	$470,186	$2,583	$140	32.3%	40.5%
2	[5, 10)	292	$199,136	$682	$50	13.7%	54.1%
3	[10, 20)	247	$319,359	$1,293	$25	21.9%	76.1%
4	[20, ∞)	407	$348,883	$857	$30	23.9%	100.0%

Code

fig = px.bar(by_band, x="distance_band", y="dollars", color_discrete_sequence=[BLUE])
fig.update_layout(
    showlegend=False,
    xaxis_title="Distance from Hubbard Hall (miles)",
    yaxis_title="Donated $",
)
fig.update_yaxes(tickprefix="$", tickformat=",.0f")
fig.update_traces(hovertemplate="$%{y:,.0f}<extra></extra>")
style(fig).show()

Figure 2.1: Donated dollars by distance band (miles)

The tightest band includes some approximate (ZIP-centroid) placements — its centroid share, to read the 1-mile count in context:

Code

prec = (
    donors.groupby("distance_band", observed=True)["geo_precision"]
    .apply(lambda s: (s == "centroid").mean())
    .to_frame("centroid_share")
)
prec.style.format({"centroid_share": pct})

	centroid_share
distance_band
[0, 1)	12.8%
[1, 5)	1.6%
[5, 10)	9.9%
[10, 20)	15.0%
[20, ∞)	14.7%

2.1 Has donor geography changed over time?

The same dollars, split into roughly five-year periods and shown as shares within each period, so changes in the mix are visible regardless of overall growth. (This uses gift-level dates, unlike the lifetime household totals above.)

Code

from hh.clean.accounts import clean_accounts
from hh.clean.donations import clean_donations
from hh.analytics.donors import succeeded_individual_gifts

PERIODS = ["2013–2015", "2016–2020", "2021–2026"]
BANDS = ["[0, 1)", "[1, 5)", "[5, 10)", "[10, 20)", "[20, ∞)"]
BAND_COLORS = dict(zip(BANDS, ["#b7d3f6", "#6da7ec", "#3987e5", "#184f95", "#0d366b"]))

accounts_geo = read_parquet("processed", "accounts_geocoded.parquet")
gifts = succeeded_individual_gifts(clean_donations(accounts=clean_accounts()))
gifts["distance_band"] = gifts["account_id"].map(
    accounts_geo.set_index("account_id")["distance_band"]
)
gifts["period"] = pd.cut(
    gifts["donation_date"].dt.year, bins=[2010, 2015, 2020, 2026], labels=PERIODS
)
gd = (
    gifts.dropna(subset=["distance_band", "period"])
    .groupby(["period", "distance_band"], observed=True)["donation_amount"]
    .sum()
    .reset_index()
)
gd["share"] = gd.groupby("period", observed=True)["donation_amount"].transform(
    lambda s: s / s.sum()
)
w10 = {
    p: f"{gd[(gd['period'] == p) & (gd['distance_band'].isin(BANDS[:3]))]['share'].sum():.0%}"
    for p in PERIODS
}

Code

fig = px.bar(
    gd, x="period", y="share", color="distance_band",
    color_discrete_map=BAND_COLORS, category_orders={"distance_band": BANDS},
)
fig.update_layout(
    xaxis_title="Period", yaxis_title="Share of donated $", legend_title="Distance band"
)
fig.update_yaxes(tickformat=".0%")
style(fig).show()

Figure 2.2: Donor-dollar share by distance band, by period

The within-10-mile share of donated dollars was 53% in 2013–2015, 55% in 2016–2020, and 54% in 2021–2026.