Payments Forecasting

Forecast gross payment volume (GPV) across merchants using panel data with 44 entities tracked over 41 time periods.

Dataset Exploration

The Payments dataset tracks financial transaction patterns and payment processing metrics. Features include currency type, territory, payment provider, and payment method flags. The target variable (GPV) represents gross payment volume.

Panel data structure: 44 merchants tracked over 41 time periods (1,804 total rows, 9 columns).

Fit Data (first 5 rows)

id	date	gpv	currency_eur	territory_FR	territory_IT	pprovider_stripe	paymenttype_link	paymenttype_sepa_debit
id4	44621	332804.72	false	false	false	true	false	false
id4	44652	25440.27	false	false	false	true	false	false
id4	44805	33363.04	false	false	false	true	false	false
id4	44682	10024.33	false	false	false	true	false	false
id4	44896	1950.75	false	false	false	true	false	false

Forecast Data (first 3 rows)

The forecast file covers future periods without the gpv target column.

id	date	currency_eur	territory_FR	territory_IT	pprovider_stripe	paymenttype_link	paymenttype_sepa_debit
id0	45809	false	false	false	true	false	false
id0	45839	false	false	false	true	false	false
id0	45870	false	false	false	true	false	false

Download sample datasets from GitHub

Code Walkthrough

Step 1: Initialize

Python

import pandas as pd
from datfid import DATFIDClient

client = DATFIDClient(token="your_DATFID_token")

Step 2: Fit the Model

Python

url_fit = "https://raw.githubusercontent.com/datfid-valeriidashuk/sample-datasets/main/Payments.xlsx"
df = pd.read_excel(url_fit)

result = client.fit_model(
    df=df,
    id_col="id",
    time_col="date",
    y="gpv",
    current_features="all",
    filter_by_significance=True
)

Step 3: Forecast

Python

url_forecast = "https://raw.githubusercontent.com/datfid-valeriidashuk/sample-datasets/main/Payments_forecast.xlsx"
df_forecast = pd.read_excel(url_forecast)

forecast = client.forecast_model(df_forecast=df_forecast)

Analysis Results (Model Fit)

Formula

gpv ~ α1*Intercept + α2*paymenttype_sepa_debit + α3*paymenttype_link + α4*territory_IT + α5*territory_FR

Alpha Estimates (Time-Invariant)

All features in this dataset are time-invariant (entity characteristics), so there are no time-varying betas.

Variable	Estimate	T-stat	Interpretation
Intercept	12,156.0	9.8	Baseline gross payment volume per entity (~12,156).
paymenttype_sepa_debit	-1,346.6	0.205	SEPA-debit payments differ from the reference payment type by ~-1,346.6 in gross volume, but not statistically significantly.
paymenttype_link	-12,073.1	2.6	Link payments run ~12,073.1 below the reference payment type.
territory_IT	-11,663.8	4.6	Italian territory runs ~11,663.8 below the reference territory.
territory_FR	-9,513.5	3.8	French territory runs ~9,513.5 below the reference territory.

Model Performance

0.018

R² Overall

0.114

R² Between

R² Within

12,577.3

MAE

1.57e9

MSE

The Payments dataset has high merchant-level variance and limited features. Adding time-varying features or lagged variables would likely improve model fit.

Try it yourself: Select "Payments" in the Free Playground.