This report analyzes the TidyTuesday 2021-03-02 release on Super Bowl Ads — 247 rows after cleaning and merge. Which Super Bowl spots won the post-game internet?
Five charts track View count across time, category, and named entities — trend, leaders, distribution, tiers, and relationships. Where companion files exist in the repo, they are joined before analysis so reception, geography, or metadata columns are not left on the table.
FAST FACTS
DATASET CONTEXT
The source is the TidyTuesday release from 2021-03-02 (R for Data Science community). This working file contains 247 rows and 26 columns after merging all available CSV/XLSX tables in the week folder.
Charts are exported as Plotly JSON with PNG fallbacks. Medians are used for robustness where distributions skew. Index-style fields (row numbers, sequential IDs) are excluded from metric selection.
How to read this report: start with the chart caption, then ask what the metric actually means, what a non-expert should notice first, and what an expert would challenge in the source. The goal is not to memorize every number; it is to leave with a sharper question than the one you arrived with.
Reader path: if you are new to the topic, treat each chart as a guided tour of one question: who leads, how concentrated the field is, what changes over time, and where the outliers sit. If you already know the domain, use the same charts as a challenge: check whether the metric is the right proxy, whether the source omits an important population, and whether the headline survives the limitations section.
CHART 1 — TREND
Median view count is falling from 111,814 in the opening period to 33,766 at the close.
Annual medians filter one-off spikes so the structural slope — not viral outliers — drives the story.
CHART 2 — LEADERS
NFL leads at 403,641 — 46,661 marks the median among the top dozen.
Head-of-field concentration is where quality, scale, or brand visibly separates from the pack.
CHART 3 — DISTRIBUTION
Category boxes reveal whether view count consensus is shared or contested across tiers.
Wide whiskers flag segments where outliers — not averages — drive reputation.
CHART 4 — CONCENTRATION
The top 5 brand entries account for 87% of the aggregate view count.
Steep Pareto curves mean a small head drives most of the signal — the long tail is noise until it isn't.
SUPPLEMENT — RELATIONSHIP
Joint plot of view count and like count surfaces clusters the averages erase.
Bubble size tracks repeat presence — outliers are archetypes, not noise.
LIMITATIONS
Community-cleaned TidyTuesday snapshots are not live APIs. Missing values, spelling variants, and week-of-export coverage limits apply. Merged tables may fan out or duplicate rows when join keys are imperfect.
Findings describe the file on hand — treat them as structural signals about Super Bowl Ads, not exhaustive truth about the full domain.
CONCLUSION
Read as a teaching map, Super Bowl Ads shows why one metric is rarely enough: leaders, tails, trends, and relationships each answer a different question about view count.
The best reading is modest: use the chart to sharpen the question, then check the source and limits before turning it into a claim.
REFERENCES
Data Science Learning Community. (2021). TidyTuesday: Super Bowl Ads. https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2021/2021-03-02/youtube.csv
EDITOR'S NOTE
Artometrics data report from the TidyTuesday research pipeline. Charts and aggregates are reproducible from the embedded exhibits and public source files.