Py.Cafe

# Streaming Services — Customer Survey 2026

A self-contained Vizro dashboard surfacing customer-survey insights for
seven major streaming services (Netflix, Disney+, HBO Max, Prime Video,
Apple TV+, Paramount+, Crunchyroll). Two interactive tabs, ten global
filters, a green & yellow visual identity.

> **Note on data.** The respondent-level CSV in `data/` is fully
> synthetic — generated deterministically by `scripts/build_mock_data.py`
> with a fixed seed. The dashboard is intended as a demonstration /
> design preview, not a real research artefact. Replace the CSV with
> your own respondent-level data (matching the schema in
> `aggregators.py`) to point it at real numbers.

## Run locally

```bash
# 1) Generate the synthetic respondent dataset (only needed once;
#    the CSV + JSON outputs are committed alongside the source).
/path/to/venv/bin/python MockupDashboard/scripts/build_mock_data.py

# 2) Boot the live Vizro app (port 8072 by default; override with PORT=…).
/path/to/venv/bin/python MockupDashboard/app.py
```

Then open http://127.0.0.1:8072/ in a browser.

## Deploy to py.cafe

py.cafe runs Vizro apps natively. To deploy:

1. Open <https://py.cafe/> and create a new project, choosing the
   **Vizro** template.
2. In the side panel, upload **every file in this folder** preserving
   the relative structure (`app.py` at the project root, plus the
   `dashboard/`, `data/`, `assets/` subfolders, plus `theme.py`,
   `aggregators.py` and `requirements.txt`). The simplest way is to
   drag-and-drop this folder's contents straight into py.cafe's file
   tree.
3. Make sure the entry file is set to `app.py`.
4. Save / Run — py.cafe installs `requirements.txt` and boots the app.

`app.py` already detects py.cafe (or any environment that exposes a
`PORT` env var) and switches to deploy mode automatically: binds to
`0.0.0.0`, disables Flask debug, disables the reloader. It also exposes
a WSGI `server` handle, so any standard PaaS (Heroku, Railway, Fly,
Render, etc.) can run it via `gunicorn app:server`.

### What to upload

Required:

```
app.py
requirements.txt
theme.py
aggregators.py
dashboard/__init__.py
dashboard/wording.py
dashboard/chart_layout.py
dashboard/chart_logic.py
dashboard/layout.py
data/respondent_level.csv
data/filter_options.json
assets/_shell.css
assets/_charts.css
assets/_menu.js
```

Optional / not needed at runtime:

```
scripts/build_mock_data.py     # only needed if you regenerate data
README.md
```

### File-size note

`data/respondent_level.csv` is ~503 KB (2,000 rows × 71 columns). All
other files are < 50 KB. Total project size is well under typical
py.cafe per-file and per-project limits.

## Tabs

- **1. Drivers of streaming choice** — heatmap of reasons × services
  with a weighted-base bubble row above each column. Three sub-tabs:
  - A. Why subscribers picked their primary service over their
       previous one (positive / green ramp)
  - B. Why subscribers' primary service is the best fit for them
       (positive / green ramp)
  - C. Why respondents would not subscribe to a service
       (negative / yellow → amber ramp)
- **2. Overall satisfaction by streaming service** — per-service stacked
  distribution bars (5 buckets: Not at all → Very satisfied) with N and
  weighted-average bubbles per row.

## Data schema

- 7 streaming services: Netflix, Disney+, HBO Max, Prime Video, Apple
  TV+, Paramount+, Crunchyroll.
- 2,000 synthetic respondents (deterministic, seed = 42).
- 10 global filter columns: gender, generation, country, income,
  monthly streaming spend, primary device, household type, employment,
  education, habitat.

To regenerate the dataset (e.g. change service-share weights or sample
size), edit `scripts/build_mock_data.py` and rerun it. The seed is
fixed so the data is reproducible.

## Project layout

```
MockupDashboard/
  app.py                       # Vizro entry point (run locally + on py.cafe)
  requirements.txt             # vizro / dash / plotly / pandas / numpy
  theme.py                     # Plotly green/yellow template + apply_theme()
  aggregators.py               # Respondent-level → tab-shaped aggregates
  dashboard/
    __init__.py
    wording.py                 # Every visible string + service & reason labels
    chart_layout.py            # Margins, legend SVGs, swatch helpers, colour cells
    chart_logic.py             # Tab 1 heatmap + Tab 2 stacked-bar figure builders
    layout.py                  # Vizro page tree + 10 filters + @capture wrappers
  data/
    respondent_level.csv       # 2,000 synthetic rows × 71 cols
    filter_options.json        # Ordered options for each filter dropdown
  assets/
    _shell.css                 # Banner + sidebar + hamburger drawer
    _charts.css                # Chart container polish
    _menu.js                   # Hamburger drawer behaviour
  scripts/
    build_mock_data.py         # Deterministic synthetic-data generator
  README.md
```