shadcn.io is not affiliated with official shadcn/ui
Insight Metrics Reviews Block for React
Three insights worked out from typed café-app reviews, each a sentence with a two-bar evidence strip and a button that opens a Sheet listing the reviews behind it.
Installation
Overview
A set of review insights for React with Base UI, for Mosshouse, a fictional neighbourhood café and its order-ahead app. A header counts the 32 typed reviews (8 low, 1 to 3 stars; 24 high, 4 and 5 stars) and says that topics are found by keyword and each sentence is a template filled in with counts. Under it, three cells with overshoot each hold a topic tag, a sentence such as "Pickup wait comes up in 5 of 8 low reviews and 2 of 24 high ones", a two-bar evidence strip (the share of low reviews and of high reviews that mention the topic, each with its count) and a button, "See the 7 reviews", that opens a Sheet with every review behind the sentence. The sentence's order follows whichever group raises the topic more, so a topic that happy reviewers mention more reads the other way round. Nothing is generated or scored: the topics, their keywords and every count are in the file, and the figures are computed from the reviews. It's built from Sheet, Badge, Button, Rating and Guides, with CSS transitions only.
FAQ
How do I use my own reviews and topics?
Replace REVIEWS and TOPICS at the top of the file. A review has an id, name, rating from 1 to 5, an ISO date, a title and a body; a topic has an id, a label and a list of whole lower-case keywords. A review mentions a topic when its title or text contains one of the words. The counts, the shares, each sentence and the list in the Sheet are all computed from those two arrays, so adding a topic adds a cell.
Is this analysis or generated text?
Neither. The sentences are fixed templates filled in with counts, and the topics are matched by keyword, so they can miss a review that says the same thing in other words. That is why the Sheet lists the matches and the words used: you can check every number against the reviews. Nothing here is a model's reading of the text.
How does the Sheet work?
"See the N reviews" opens the catalog Sheet with the matching reviews, newest first, each with its stars, title, text, reviewer and date, and a line saying how many were low and how many high. Close it with the close button, Escape or a click outside, and focus returns to the button that opened it. It slides in from the end edge on a wide screen and up from the bottom on a phone.
Related
- Pro
Keyword Mentions Reviews
Insights and trendsA topic table for sofa reviews: mentions found by keywords, a share bar and the average rating of the reviews that mention each topic against the overall average, sortable, with the matching reviews under it.
- Pro
Trend Overview Reviews
Insights and trendsA router's rating over twelve months as a line with monthly review-volume bars on one axis, a 6 or 12 month switch, a month Slider and a closing band reading that month's average, count and most helpful review.
- Pro
Analytics Summary Reviews
Insights and trendsSix products of one outdoor shop ranked by their reviews: a sortable table of average, count, a five-segment ratings bar, the 30-day change and the answered share, all worked out from typed reviews.
- Pro
Typed Summary Reviews
Insights and trendsA hand-written, labelled summary of 36 earbud reviews: two sentences that each end in a count opening a popover of three snippets, the average rating and five themes with mention counts.
- Pro
Tag Cloud Reviews
Insights and trendsA word cloud for coffee-bean reviews in a framed figure: eighteen words sized by how often the reviews say them, and a keyword-in-context list that shows every mention of the chosen word with who said it.
- Pro
Sentiment Badges Reviews
Filter and searchSix meal-delivery reviews each marked with a tone badge from a printed rule, with a drawn shape and a word, and a chip filter by tone with counts.