Sample Dataset Gallery
Free, ready-to-use CSV datasets — the same ones we chart in our Dataset de jour series. Download one, drop it into Power BI, and point LLM AI Charts at it to see what plain-English prompts can do.
Fitness Tracker — Daily Steps, Sleep & Heart Rate
Daily wearable-device activity for a cohort of users: step counts, calories burned, average heart rate and sleep hours, with weekday and per-user grain. A rich mix of a date, a category (user), and several measures — good for time-series, per-category comparison, and tabular scorecards.
Airline Flight Delays
Flight-level delay minutes by origin, destination and airline, plus distance, departure hour and a cancelled flag. Route-and-carrier structure that suits flow/network views, delay heatmaps, and hour-of-day distributions.
Business Locations Directory
A multi-national retail chain's store directory — one row per location with a full mailing address (address line, city, state/province, country, postal code) plus sales region, store format, annual revenue, headcount, floor area and an opened date. Real, coherent geography across 30 cities and 11 countries — built for maps, revenue-by-country/region rollups, store-format breakdowns, and address-level tables.
City 311 Service Complaints
Municipal 311 service requests by complaint type and neighborhood, with days-to-resolve, resolution cost, priority and a same-week-resolved flag. Categorical-heavy with an ordinal priority — a natural fit for ranked bars, heatmaps, and resolution-rate breakdowns.
Company Org Chart with Headcount
Twelve departments nested inside four divisions, each with its headcount. Two levels and one measure is the shape of an org tree, and the headcount lets each box carry a bar that compares it with the other boxes at its level.
Consulting Engagement Billings
Six years of individual consulting billing events — a plain date column (sparse, business days with gaps), a two-level Practice → Service Line hierarchy, and a billing amount per engagement. One practice launches mid-series and another spikes every Q4, so time-grain choices matter: raw dates demand aggregation to weeks, months or quarters before charting. Built for date-bucketing exercises: animated treemaps, calendar heatmaps, small multiples and seasonal breakdowns.
Country Development Indicators (Illustrative)
Forty-five countries with seven indicators each: human development index, life expectancy, GDP per capita, mean years of schooling, internet users, CO2 per capita and a Gini index. The values are generated to look plausible and to move together; they are not official statistics, and a country's row should not be read as that country's real figures. Two numeric measures per country make it a sample for bivariate choropleths, scatter plots and correlation views.
COVID Cases & Hospitalizations by State
Daily new COVID cases and hospitalizations by U.S. state. A date + geographic category + two measures across many states — a classic small-multiple time series and state-comparison set.
Daily Stock Prices (OHLCV)
Daily open/high/low/close and volume for multiple tickers. Textbook OHLCV time series — for candlestick and line charts, volume overlays, and multi-ticker comparison.
Department Budget vs Actual
Eight departments, each with an actual spend and the budget it was held against. The rows carry the cases a variance chart has to handle: departments well over and well under budget, one exactly on budget, and one with a budget of zero. One category and two measures is the smallest useful shape for a variance chart, a bullet chart or a paired bar.
Divisional Revenue by Year
Twelve years of revenue for four divisions, each split into four product categories — a 2-level hierarchy observed annually, with realistic growth, decline and shock patterns (a digital boom, a 2020 retail slump, print's slow fade). Built for hierarchy-over-time views: animated treemaps, streamgraphs, stacked areas and bump charts.
Earthquake Locations & Magnitude
Point events with latitude, longitude, magnitude and depth. A compact lat/long + measure set for map scatter, magnitude-vs-depth relationships, and density views.
Global Revenue Expansion — 20 Years of Country Rollout
Twenty years (2006–2025) of revenue and customer counts for a company opening one market at a time: the United States alone for two years, then Canada, then Mexico / the UK / Japan, then western and northern Europe, then the rest of Asia and Australia, and finally New Zealand and three African markets — 1 country on the map in 2006, 35 by 2022. Countries contribute no rows at all before they open, and each one starts small and compounds from its own entry year, so an animated choropleth genuinely fills in over time rather than just changing shade. Country names and ISO-3 codes both resolve; grain is Year × Country × product line with revenue and customer count, left unaggregated so the engine picks the summarization.
Household Expense Breakdown
Monthly household spend by category. A clean date + category + measure triple — the simplest starting shape for stacked area, part-of-whole, and category trend charts.
KPI Attainment vs Target
A single row: a Q3 service objective with an attainment of 87.4 against a target of 95, and a capacity utilization figure of 0.734. The smallest shape there is, one value with the ceiling it is measured against, which is exactly what a gauge, a linear gauge, a thermometer or a progress ring draws.
Loan Default Predictors
Loan records with credit score, annual income, debt-to-income, purpose, amount and an outcome (defaulted) flag. Several continuous measures plus a binary target — ideal for scatter/bubble, distribution, and default-rate-by-segment analysis.
Monthly Revenue at Nine Growth Rates
Two years of monthly revenue, January 2024 to December 2025, repeated for each of nine scenario growth rates from -2% to +6%. Each row is a month, a growth rate and the revenue that rate produces, so all nine scenarios can be drawn on one plot and stepped through.
Monthly Signups with a Target
Thirty months of signups, January 2023 to June 2025, beside a constant target of 13,500. One measure over time with a fixed goal, which is the starting point for a what-if projection: choose a growth rate and a horizon and see the month the target is reached.
NBA Player Season Stats
Per-player season averages — points, rebounds, assists, field-goal and three-point percentages, with team, position and age. A wide sports stat line for player comparison, position breakdowns, and shooting-efficiency scatter. Pairs with the NBA Shot Locations dataset.
NBA Shot Locations
Individual basketball shot attempts with court X/Y coordinates, made/missed outcome, quarter and team. Point data with a spatial x/y and a binary outcome — ideal for shot-chart scatter, hexbin density, and make-rate-by-zone heatmaps.
North America City Metrics — 21 Cities with Coordinates
Twenty-one cities spanning the United States, Canada, Mexico and Central America — deliberately wider than the lower 48, so a North America basemap actually earns its shape. Each row carries explicit latitude/longitude (the one gallery sample with raw coordinates rather than resolvable names), a country, a sales region, and five measures: population, revenue, store count, a satisfaction score and a growth rate that goes negative in three markets. Country to sales region to city is a real three-level hierarchy, so it suits a treemap, an icicle or a sunburst as readily as a map. Population against revenue is the interesting pair: size bubbles by one, colour by the other, or derive revenue-per-capita and see which small markets punch above their weight. This is the dataset behind the article's North America bubble map — and its star-marker variant.
Ocean Shipping Container Manifest
Container-level ocean freight bookings — revenue, weight, sail/arrival dates, voyage, vessel, load and discharge ports, carrier, commodity, customer tier and status. Very wide, high-volume logistics data for lane/route maps, on-time analysis, and revenue-by-lane breakdowns.
Order Fulfilment Process Flow
Twenty-two hand-offs between the twelve steps of an order fulfilment process, from order received to delivered, each with the team that owns the step it starts from and the number of orders that took that path. Backorders, cancellations, rework and returns loop back or leave the main line, so it is a real process rather than a straight chain. A from-step, a to-step and a team make a flowchart with a swimlane per team.
Order Value by Store, Raw Order Rows
2,623 individual orders across 205 stores, each store identified only by an opaque code, with one order value per row. Stores hold anywhere from a single order to 28, and order values are heavily right-skewed: a few very large orders against a much lower median. With that many stores there is no room for a bar per store, so it suits views of the distribution such as a beeswarm, a strip plot or a box plot, and a headline total with the spread behind it.
Patient Vital Signs
Per-visit clinical vitals — temperature, systolic/diastolic blood pressure, heart rate and oxygen saturation, keyed by patient and visit. Wide, multi-measure health data for per-patient trends, distribution/box plots, and vitals correlation.
Product Release Timeline
Twenty-four release milestones from July 2024 to May 2026, each with a date, a name and one of three tracks (Platform, Mobile and Data). Two milestones share a date. A date beside an event name is all a timeline needs, and the tracks run in parallel.
Real Estate Listings
Property listings by neighborhood and type — square footage, bedrooms, bathrooms, age and list price. A wide mix of categoricals and measures — good for price-vs-size scatter, price-by-neighborhood, and bed/bath breakdowns.
Restaurant Orders & Tips
Individual restaurant orders by date, weekday, party size and meal period, with subtotal, tip percent and tip amount. High row count with a date and several measures — good for tip-rate trends, weekday/meal-period seasonality, and party-size effects.
Retail Store Daily Sales (2 Years)
Two years of daily store sales by region, store and channel — units sold, revenue and returns. A large date + multi-category + measures set — great for trend lines, region/channel breakdowns, and return-rate analysis.
Revenue Mix by Account
Twenty-two accounts in three segments (Enterprise, Mid-market and SMB), each with the share of its revenue that came through direct, partner and self-serve channels, plus its annual revenue. The three shares add up to 100 on every row, to within rounding, which makes it a natural ternary plot: each account is one point in the triangle and the segments show up as clusters. It also suits 100% stacked bars.
Ride-Hailing Demand by Hour
Hourly ride demand by neighborhood — rides requested vs completed, with weekday and hour-of-day. A date/time + category + paired measures — good for demand heatmaps, fulfillment-rate trends, and peak-hour analysis.
SaaS Subscription Metrics by Plan
Eighteen months of subscription metrics, January 2024 to June 2025, for four plans (Starter, Pro, Business and Enterprise): active subscriptions, monthly recurring revenue, churn rate and new signups. A month, a category and four measures at a small row count, well suited to chart types that move entities across two axes over time, such as the motion bubble chart.
Solar Panel Energy Output
Daily per-panel solar generation — energy produced, peak sun hours and temperature. A date + panel id + measures — ideal for output trends, sun-hours-vs-energy relationships, and per-panel comparison.
Store Revenue by Country, 15 Years
Monthly revenue by store format for 34 countries over fifteen years, January 2011 to December 2025: 14,400 rows, with each country running one to four of five formats (Airport, Downtown, Flagship, Online and Outlet). Every country appears under one spelling, so a country ranking such as an animated bar-chart race draws one bar per country rather than a bar for each variant of its name.
Support Cost by Team, Year over Year
Four teams with their support cost in 2023 and again in 2024, one column per year. Costs rise for two teams and fall for two, and nothing in the data says which direction is good, so a delta KPI card has both directions to show without judging either. A compact wide-format shape for year-over-year comparison, slope charts and dumbbells.
Team Quota Attainment
Seven regional teams with what each attained against its quota. The awkward cases are in there on purpose: one team far past its quota, one exactly on it, one just over it, and one with an attainment but no quota at all. Suits gauges, linear gauges, thermometers and progress rings, where a team with no target has to be handled rather than drawn as zero.
U.S. Orders by ZIP Code
Ten thousand order rows keyed by 5-digit U.S. ZIP, with city, state, sales channel, order count and revenue. The 8,354 distinct ZIPs roll up to 853 ZIP-3 prefixes across all 51 state codes — enough coverage to actually fill a USA-by-ZIP-3 choropleth rather than leave it mostly blank. Three separate place columns (ZIP, city, state) and two measures make it the one sample that drives every U.S. map type: a ZIP-3 choropleth, a by-state choropleth, and a North America bubble map with orders sizing the bubbles and revenue colouring them. ZIPs keep their leading zeros here; if your import strips them (Excel and Power BI both will, given the chance), the visual pads them back rather than misplacing the 0xxxx block.
Urban Air-Quality Monitoring
Daily pollutant concentrations across several cities with an AQI category label. A date + two categoricals (city, pollutant) + a measure — built for small-multiple time series, city comparisons, and category-banded trends.
US State Store Metrics
All fifty states and the District of Columbia, one row each, with total revenue, total orders and a return rate. Revenue runs from about half a million in the smallest state to over thirty million in the largest, and the return rate from 2.1 to 12, so revenue against return rate gives a bivariate map places where the two disagree. Also a fit for a hex-tile cartogram, where every state gets the same shape and the small ones stay readable.
Wine Chemistry & Ratings
Physicochemical wine attributes — alcohol, acidity, residual sugar, sulfates, pH, tannins and density — with variety, region and a quality rating. A very wide numeric feature set for correlation heatmaps, scatter matrices, and rating-driver analysis.
World City Sales — 117 Cities Across 55 Countries
One row per city — 117 of them across 55 countries, each with its country name AND ISO-2 code, world region, a city tier (Capital / Primary / Secondary), order count, revenue, and average order value. The city names alone resolve to real coordinates — no latitude/longitude columns anywhere — which is exactly the lookup a world bubble map needs, and the country columns roll the same rows up into a world choropleth. Orders and revenue give the bubbles independent size and colour; AvgOrderValue is a per-row rate, so treat it as an average, not a sum.