Business Analytics in Action: From Model to App

The Mancino dataset

Mancino Market is a synthetic grocery retailer: fictional daily unit sales for five stores and fifteen product categories, used throughout the course. Source: mancino_daily_sales.csv (v1_2025-11-01 snapshot, bundled offline; see data/SOURCE.md).

Loading dataset…

Variable dictionary — what the model predicts and what it uses

Field	Meaning	Role
units_sold	Daily units for one store and category	Prediction target
store_id	Store identifier (1–5), treated as text	Model input (categorical)
category	Product category	Model input (categorical)
weekday	Derived from date in Python	Model input (categorical)
day_of_year	Derived from date	Model input (numeric)
date	Selling date	Derives weekday and day_of_year; sets the date splits
revenue	Realized revenue	Visible only — contains the outcome, never an input
avg_price	Observed daily average price	Visible only — not known before the selling day
day_of_week	Weekday in the source file	Visible only — weekday is recomputed from date

Browse the data

These filters change the view only; they never change model fitting or evaluation data.

Store (view filter) Category (view filter) Search table text (store, category or weekday)

Total units sold per day (across the stores and categories in view)

Choose one model, train it, read its results

Model Parameter

How it predicts:

No results yet for this configuration. Click “Train model” to fit it on the development dates and see its four test metrics.

Code from class

Read directly from models.py — the dashboard has no second copy of the model code.

The Mancino dataset

Variable dictionary — what the model predicts and what it uses

Browse the data

Total units sold per day (across the stores and categories in view)

Choose one model, train it, read its results

Test the model with a typed observation