Datasets

Pinball bundles two classic datasets used throughout the quantile regression literature. Both are returned as sklearn.utils.Bunch objects for compatibility with sklearn workflows.

---

load_engel

from pinball.datasets import load_engel

Description

Ernst Engel's 1857 food expenditure data. 235 Belgian working-class households with a single predictor (income) predicting food expenditure. This is the canonical dataset for illustrating quantile regression — it appears in virtually every tutorial and textbook.

Returns

Key	Type	Shape	Description
data	ndarray	(235, 1)	Household income
target	ndarray	(235,)	Food expenditure
feature_names	list[str]	—	["income"]
DESCR	str	—	Human-readable description

Example

from pinball.datasets import load_engel
from pinball import QuantileRegressor

engel = load_engel()
X, y = engel.data, engel.target

model = QuantileRegressor(tau=[0.1, 0.25, 0.5 , 0.75, 0.9])
model.fit(X, y)

# The slopes diverge: wealthier households show more variability
# in food spending — a textbook example of heteroscedasticity.
print(model.coef_)

Why this dataset?

Engel's law states that the proportion of income spent on food decreases as income rises. Quantile regression reveals an additional insight: the variability of food spending also increases with income. The slope at the 90th percentile is much steeper than at the 10th — richer households have more diverse food budgets.

---

load_barro

from pinball.datasets import load_barro

Description

Cross-country economic growth data from Barro (1991) and Barro & Lee (1994). 161 countries with 13 predictor variables and net GDP growth as the target. This multi-predictor dataset is useful for demonstrating quantile regression with several covariates.

Returns

Key	Type	Shape	Description
data	ndarray	(161, 13)	Predictor variables
target	ndarray	(161,)	Net GDP growth rate
feature_names	list[str]	—	See table below
DESCR	str	—	Human-readable description

Feature Names

Column	Name	Description
0	lgdp2	Log initial GDP per capita
1	mse2	Male secondary school enrolment
2	fse2	Female secondary school enrolment
3	fhe2	Female higher education
4	mhe2	Male higher education
5	lexp2	Life expectancy
6	lintr2	Log investment rate
7	gedy2	Government education / GDP
8	Iy2	Investment / GDP
9	gcony2	Government consumption / GDP
10	lblakp2	Log black-market premium
11	pol2	Political instability
12	ttrad2	Terms of trade

Example

from pinball.datasets import load_barro
from pinball import QuantileRegressor

barro = load_barro()
X, y = barro.data, barro.target

# Median regression
model = QuantileRegressor(tau=0.5, method="fn")
model.fit(X, y)

# Which predictors matter at the lower tail?
model_low = QuantileRegressor(tau=0.1)
model_low.fit(X, y)
print(dict(zip(barro.feature_names, model_low.coef_)))

References

1. Barro, R. (1991). "Economic growth in a cross section of countries." Quarterly Journal of Economics 106(2): 407–443.
2. Barro, R. and Lee, J.-W. (1994). "Sources of economic growth." Carnegie-Rochester Conference Series on Public Policy 40: 1–46.