Movie Financial Performance Dashboard

A personal, non-commercial Power BI portfolio project that turns a movie spreadsheet exercise into a reproducible analytics pipeline and semantic model. The project began with a multi-page starter Excel workbook with only ~500 movies, restructured its data into fact and dimension tables, and then expanded and enriched the dataset with the The Movie Database (TMDb) API. The current analytical sample contains 2,371 movies released from 2010 to 2025 and 7,007 credited people.

Dashboard

Power BI file: Download MoviesDashboard.pbix

The overview summarizes movie count, revenue, profit and ROI, with comparisons across genres, directors or actors, release months, budgets and box-office revenue. All visuals respond to the release-year filter.

Report Pages

Genre Analysis

The genre page compares movie volume, revenue, profit and ROI. Multi-select genre controls support comparisons across release timing, financial scale and average return.

Director Details

The director page combines financial KPIs with a selected director's profile, biography and filmography. The movie table includes budget, revenue, profit, ROI and rating.

Cast Details

The cast page provides the same person-level experience for actors, linking a profile and biography to the movies and financial results available in the curated dataset.

Movie Tooltips

Hover tooltips add a compact movie snapshot with title, tagline, overview and poster without requiring a separate report page.

Data Journey

block-beta
    columns 7
    space A["Kaggle starter<br/>Excel"] space B["Cleaned and<br/>restructured tables"] space C["Fact-dimension<br/>semantic model"] space
    space space space space space space space
    space F["Power BI<br/>report"] space E["TMDb person-detail<br/>enrichment"] space D["TMDb movie expansion<br/>and enrichment"] space

    A --> B
    B --> C
    C --> D
    D --> E
    E --> F

    classDef phase font-size:18px
    class A,B,C,D,E,F phase

Loading

The project was developed in phases:

1. The starter workbook was profiled and reshaped into staging, fact, dimension and bridge tables.
2. TMDb became the primary source for a larger, cleaner movie sample, including movie details, credits, genres, posters and external identifiers.
3. Person details were added separately so biographies, birth information and profile images could enrich the Director and Cast pages without widening the core person dimension.
4. The generated CSV tables were loaded into a Power BI semantic model and used to build the interactive report.

Semantic Model

The model separates movie performance measures from descriptive dimensions. Bridge tables resolve the many-to-many relationships between movies and genres and between movies and credited people. Optional person details connect to dim_person through person_key.

Power BI loads these model tables from data/processed/:

• dim_movie.csv
• fact_movie_performance.csv
• dim_genre.csv
• bridge_movie_genre.csv
• dim_person.csv
• dim_person_detail.csv
• bridge_movie_credit.csv
• dim_date.csv

movie_identity_map.csv is retained as an optional matching and QA reference.

Current Dataset

The generated dataset is a curated TMDb sample covering releases from 2010-2025.

• Movies: 2,371
• Target depth: 150 clean, revenue-sorted movies per year
• Current exception: 2020 contains 121 movies after cleaning
• Deduplication: one movie per tmdb_id
• Minimum budget: $1M
• Required movie fields: title, release date, runtime, revenue, budget, genre, director and cast
• Enriched people: 7,007

This is a clean analytical sample rather than a complete movie catalog.

Pipeline

TMDb responses are cached locally, normalized into staging CSVs and transformed into stable Power BI tables. The raw workbook and API cache remain local. More detail is available in source notes and the data dictionary.

Create the environment and add a TMDb bearer token to a local .env file:

mamba env create -f environment.yml
conda activate movie-dashboard

TMDB_BEARER_TOKEN=your_token_here

Run the TMDb-first pipeline:

python -m src.load_tmdb_seed
python -m src.build_model
python -m src.enrich_tmdb
python -m src.build_model
python -m src.enrich_tmdb_people
python -m src.validate

The original starter-workbook loader remains available for comparison or backfill work:

python -m src.load_seed
python -m src.build_model
python -m src.validate

Project Structure

assets/          Power BI file, report screenshots and logo assets
data/manual/     reviewed corrections and TMDb match overrides
data/processed/  generated CSV tables for Power BI
data/raw/        local-only workbook and TMDb API cache
docs/            data dictionary and source notes
src/             ingestion, enrichment, modeling and validation code
tests/           focused pipeline tests

Reproducibility Notes

• .env, data/raw/ and the TMDb API cache are not committed.
• Generated publishable CSVs are stored in data/processed/.
• The raw starter workbook is excluded because redistribution rights are unclear.
• The PBIX file is provided as a report artifact; the Python pipeline is the reproducible source for the underlying dataset.