The target

Intermediate Predictive Analytics in Python

Nele Verbiest Ph. D.

Senior Data Scientist @PythonPredictions

Target definition

Target timeline (1)

Target timeline (2)

Target timeline (3)

Defining the target in Python

unsubscribe_2017[:5]

[90112, 65537, 24577, 8196, 73737]

baseteable.head()

   donor_id
0     65537
1     65538
2         4
3     98328
4     65564

basetable["target"] = pd.Series([1 if donor_id in unsubscribe_2017 else 0 \
for donor_id in basetable["donor_id"]])

Defining an aggregated target in Python

print(gifts.head(2))

   donor_id
0     65537
1     65538

# Target period
start_target = datetime(year = 2017, month = 1, day = 1)
end_target = datetime(year = 2018, month = 1, day = 1)

# Select target period donations
gifts_target = gifts[(gifts["date"]>=start_target) & (gifts["date"]<end_target)]

# Group and sum donations by donor
gifts_target_byid = gifts_target.groupby("id")["amount"].sum().reset_index()

# Derive targets and add to basetable
targets = list(gifts_target_byid["id"][gifts_target_byid["amount"]>500])

basetable["target"] = pd.Series([1 if donor_id in targets else 0 for donor_id in basetable["donor_id"]])

The basetable

print(basetable.head())

   donor_id  target
0     65537  0
1     65538  1
2     65539  0
3     65540  1
4     65541  0

Let's practice

Intermediate Predictive Analytics in Python