Loading...
The perfect first SQL and data project: answer three questions with SQL on 250 top-rated films, then turn the results into a polished, shareable Tableau dashboard. No installs, about 45 minutes.
Conor Robertson
July 22, 2026 · 12 min read
The questions people ask before starting their first SQL and Tableau project.
QueryCase teaches SQL through mystery cases on real datasets. Music, sports, film, and more. Your first cases are completely free.
Start your first caseFree to start. No credit card required.
Learning SQL is one thing; having a project to show for it is another. This is that project, and it makes a great first one. In about 45 minutes, using two completely free tools, you will write three SQL queries against real data, then turn the answers into a polished, interactive dashboard you can put in a portfolio and share with a link.
The data is 250 of the highest-rated films on IMDb: their ratings, runtimes, genres, directors and cast. You do not need to be good at SQL yet. Every query is written out and explained in plain English, ready to copy and paste. There is no database to set up and nothing to install, from the first query to the finished dashboard it all runs in your browser.
You finish with a live, public dashboard link for your CV, LinkedIn or portfolio.
You need two free tools. Set both up now and the rest of the project is smooth. Budget about 30 to 45 minutes in total.
This is where you write and run the SQL. Nothing to install, it runs in your browser.
You will land on the SQL editor with the IMDb tables listed down the left. That is where the three queries below run.
This is where you build the dashboard, free and entirely in your browser. No download needed.
One thing to know: it is Tableau Public, not Tableau Cloud (which is paid). Everything you publish to Public is public, which is exactly what you want for a portfolio piece, because it gives you a shareable link.
Three queries, each answering a different question and producing a file you will use in Tableau.
Let's build them.
What it answers: everything at the film level. One row per film, with its rating, vote count, runtime, the decade it belongs to, its director, its lead actor and how many genres it spans. This single table will drive most of the dashboard.
Run this in the sandbox:
SELECT
m.title,
m.year,
m.year - (m.year % 10) AS decade,
r.imdb_score,
r.vote_count,
m.runtime_min,
(SELECT STRING_AGG(p.name, ', ')
FROM directors d
JOIN people p ON p.person_id = d.person_id
WHERE d.movie_id = m.movie_id) AS director,
(SELECT p.name
FROM cast_members c
JOIN people p ON p.person_id = c.person_id
WHERE c.movie_id = m.movie_id
ORDER BY c.billing_order
LIMIT 1) AS lead_actor,
(SELECT COUNT(*) FROM movie_genres mg
WHERE mg.movie_id = m.movie_id) AS num_genres
FROM movies m
JOIN ratings r ON r.movie_id = m.movie_id
ORDER BY r.imdb_score DESC;What it does, in plain English: the main query joins movies to ratings. Then three subqueries fill in the columns that need a lookup: the director (joined through the directors bridge table), the top-billed actor (the lowest billing_order in cast_members), and a count of the film's genres. The year - (year % 10) trick turns 1994 into the decade 1990.
You get 250 rows. A few documentaries (like Senna and 20 Days in Mariupol) have no cast in the data, so their lead_actor is blank. That is real, honest data, and Tableau handles it fine.
Now click Export CSV above the results and save the file as films.csv.
What it answers: genre is the one thing a film has many of, so it needs its own table with one row per film-and-genre. That lets Tableau count films per genre and average ratings by genre without double-counting.
SELECT
m.title,
m.year,
m.year - (m.year % 10) AS decade,
g.name AS genre,
r.imdb_score,
r.vote_count
FROM movies m
JOIN ratings r ON r.movie_id = m.movie_id
JOIN movie_genres mg ON mg.movie_id = m.movie_id
JOIN genres g ON g.genre_id = mg.genre_id
ORDER BY m.title, g.name;What it does: this walks the bridge from movies to genres through the movie_genres link table. Because a film like The Dark Knight is Action, Crime and Drama, it appears three times here, once per genre. That is on purpose. You get 644 rows.
Export it as film_genres.csv.
What it answers: which actors appear again and again across these great films, and how their films rate on average. One row per actor, for anyone in at least four of the 250.
SELECT
p.name AS actor,
COUNT(*) AS film_count,
ROUND(AVG(r.imdb_score), 2) AS avg_score,
MIN(m.year) AS first_film,
MAX(m.year) AS latest_film
FROM cast_members c
JOIN people p ON p.person_id = c.person_id
JOIN movies m ON m.movie_id = c.movie_id
JOIN ratings r ON r.movie_id = m.movie_id
GROUP BY p.person_id, p.name
HAVING COUNT(*) >= 4
ORDER BY film_count DESC, avg_score DESC;What it does: it joins the cast to the people and the films, groups by actor, and counts their appearances. HAVING COUNT(*) >= 4 keeps only the regulars. The result is a tidy leaderboard: Robert De Niro and Mark Ruffalo lead with nine films each, while Morgan Freeman's films carry the highest average score.
Export it as top_actors.csv. You now have all three files.
Now you turn those three CSVs into a dashboard, all in the browser. Take it one step at a time.
Sign in to Tableau Public. In the top menu, click Create, then Web Authoring to build right in your browser. (If you would rather have more formatting control, you can download the free Tableau Public desktop app instead; it uses the same free account, and every step here works the same way.)
A Connect to Data window opens. Click the Files tab, then Upload from computer (or drag the file straight in), and choose films.csv. Tableau reads your columns automatically and drops you into the editor.
Now bring in the other two files. The important part, and the bit people get wrong: add each one as its own separate data source, not as an extra table or a union inside the films connection.
On the Data Source page (the tab at the bottom-left), click the small dropdown arrow next to the data source name and choose New Data Source. Upload film_genres.csv. Then do exactly the same again for top_actors.csv.
You will end up with three independent data sources. You can switch between them from that same dropdown, and each chart later just uses whichever one it needs.
One thing about Tableau Public
Each CSV is its own data source, so charts built from different files will not cross-filter each other automatically. For a first dashboard that is completely fine. If you want them linked later, you can relate film_genres back to films on title in Tableau's data model, but do not let that hold up version one.
Each sheet below is one new worksheet (the tabs along the bottom). Build them one at a time, then we will drop them onto a dashboard.
The four tiles at the top are each a single number on its own worksheet. Watch one build, from the films source, then repeat the same moves for the other three:
Only the field, the maths and the number format change from one tile to the next:
| Tile | Drag this field | Set aggregation | Format |
|---|---|---|---|
| Average IMDb ScoreShown above | Imdb Score | Average | 2 decimal places |
| Movie Count | Title | Count | Whole number |
| Avg Runtime | Runtime Min | Average | Whole number + " mins" |
| Total Votes | Vote Count | Sum | Thousands separator |
From the films source: Decade on Columns, count of films on Rows. The clip shows the build; the payoff is seeing which eras the canon leans on.
The most revealing chart, and still on films: Imdb Score against Vote Count, one dot per film. It shows that acclaim and popularity are not the same thing. Drop Title, Director and the score onto the Tooltip so every point tells its story on hover.
Switch to the film_genres source: average Imdb Score by Genre, sorted high to low. One row per film-genre keeps that average honest.
From top_actors: Film Count by Actor, sorted down, with Avg Score on Colour, so you see who appears most and whose films rate best at once.
Create a new Dashboard (the icon beside the worksheet tabs) and set its size. Drag your sheets on: the four KPI tiles in a row across the top, the charts in a grid below. Switch an object to Floating when you want to place it exactly; leave it Tiled and Tableau keeps everything aligned for you.
A considered-looking dashboard reads as more credible, and five quick moves do most of the work:
Design resources
Grab the IMDb logo to drop into your dashboard title, for a clean nod to the source.
Click Publish in the top bar and give your workbook a name. It saves to your Tableau Public profile and gives you a public URL. That link is the whole point: it is a live, interactive piece you can drop into a CV, a LinkedIn post or a portfolio, and anyone can open it without installing a thing. Here is the finished example from this walkthrough.
The starter dashboard is exactly that: a start. The version worth showing off, and the one that teaches you the most, is the one you push past it. Take it further:
films.csv), a runtime distribution, or average rating by decade to see whether the greats cluster in one era.A dashboard you can point to is worth ten you can only describe, and "I built this" is a far stronger thing to say in an interview than "I learned some SQL."
Post your finished dashboard on LinkedIn and tag @QueryCase. We would love to see it, and we reshare the ones we love.
You just did the full loop a data analyst does every day: shape data with SQL, decide what each query answers, and communicate it visually. The SQL is the part that compounds, and the fastest way to get fluent is to keep writing it against data you actually care about.
That is what QueryCase is built for: guided cases that teach SQL as a detective story, drills between topics, and the same sandbox you just used, ready for your next question. Open the IMDb Ratings sandbox again and see what else the data will tell you.