Quick Start

1. Install the SDK

pip install datfid

2. Get Your API Key

You need an API token to use the SDK. Getting one is free and takes about 15 minutes.

Fill out a short form and the DATFID team will respond within 15 minutes with your personal token. See Get Your API Key for details.

3. Initialize the Client

import pandas as pd
from datfid import DATFIDClient

client = DATFIDClient(token="your_DATFID_token")

Sample Datasets

Ready-to-use sample data is available on GitHub: DATFID sample-datasets repo

4. Fit a Model

Load your panel data and fit an interpretable model. DATFID will return a formula with coefficients that explain the relationships in your data.

df = pd.read_excel("your_data.xlsx")

result = client.fit_model(
    df=df,
    id_col="Product",
    time_col="Time",
    y="Revenue",
    current_features="all",
    filter_by_significance=True
)

5. Explore the Result

After fit_model() completes, the returned result object exposes all analysis outputs. In Python or Colab, type result. and press Tab to see autocomplete suggestions. Key fields:

• result.formula— the fitted model as a readable equation
• result.alpha— per-entity intercepts (time-invariant baseline for each entity)
• result.beta— feature coefficients with t-stats and p-values
• result.Performance— overall R², MAE, MSE, RMSE
• result.R2_individual— per-entity R² breakdown
• result.dropped_cols— features removed (e.g. insignificant)
• result.df— processed training DataFrame used for fitting

print(result.formula)
print(result.Performance)
print(result.alpha.head())

See Python SDK Reference for the full list of result attributes including headers_alpha, headers_beta, and errors.

Drill into a single entity

You can also index result with an entity name and chain any attribute directly — all the same fields are available, scoped to that entity only:

print(result["ind1"].alpha)        # this entity's baseline intercept
print(result["ind1"].Performance)  # R², MAE, MSE, RMSE for this entity only
print(result["ind1"].formula)      # formula with this entity's alpha substituted in
print(result["ind1"].df.head())    # training rows for this entity only

See individual entity access in the SDK reference for the full attribute list.

6. Generate Forecasts

Load a forecast file that defines which entities and time periods to predict, then run the forecast.

df_forecast = pd.read_excel("your_forecast_data.xlsx")

forecast = client.forecast_model(df_forecast=df_forecast)

Next Steps

• → Python SDK Reference — full parameter reference for fit_model() and forecast_model()
• → REST API Reference — use DATFID via HTTP endpoints
• → Use Case Examples — see real results with sample datasets