Sentiment

The Sentiment tab measures how the conversation receives the subject: the balance of positive and negative voices, the days reception ran warmest and coldest, and the direction it is moving. Its central instrument is the sentiment score, a single reading of that balance on a scale from -100 to +100. Where the Overview tab reports the score as a headline, this tab adds the time dimension, which is usually where the actionable findings are: a score of -20 matters differently depending on whether it has been stable for weeks or arrived three days ago.

The page adapts to the number of queries in the query bar. With a single query it shows the layout described in the next section. When two or more queries are active, a toggle appears above the results with three views: Compare sets the queries against each other, Aggregate combines them into one conversation, and Select focuses on one of them at a time. For the query bar itself, including syntax, time ranges, and display names, see Getting Started.

Single query (Select)

The single-query layout, which the Select view reproduces for one chosen query.

Stats bar

Sentiment is the overall sentiment level, such as Mostly Positive or Very Negative, with the numeric score beneath it. It is the one-glance verdict on how the subject is being received; the rest of the tab shows what the verdict is made of and whether it is holding.

Most Positive Day (or Week, matching the chart interval) is the completed interval with the highest sentiment score, with its date and level. It marks the conversation's best moment in the range, often traceable to a specific announcement or story. Narrowing the time range to that date shows what the conversation looks like when it goes well, which is as instructive as diagnosing the bad days. If even this best day sits at a weak level, the subject had no genuinely good day anywhere in the range.

Most Negative Day (or Week) is the interval where reception bottomed out. This is typically the first cell to act on: narrow the range to that date and read the feed to identify the trigger, then check the Emotions tab over the same days to see whether the reaction runs on anger, fear, or disappointment, which are different problems wearing the same score.

Sentiment Trend is the average change in score per interval across completed intervals, in points, labeled Improving or Declining, or Stable when there is no net movement. It adds direction to the level, and the two together say more than either alone: a Mostly Negative conversation that is improving and a Mostly Positive one that is declining are opposite situations that the level by itself would conflate.

Sentiment score

The gauge places the conversation's sentiment score on a scale from -100 to +100, with its level label beneath. The score is the balance of opinion among posts that take a side: the number of positive posts minus the number of negative ones, divided by their combined total. At +100 every opinionated post is positive, at -100 every one is negative, and near 0 the two sides are evenly matched. Neutral posts stay out of this calculation, so the score reads the direction of opinion without being diluted by indifferent posts; how much of the conversation is neutral is what the distribution chart beside the gauge shows.

Scores map to five levels used across the dashboard: Very Positive above +50, Mostly Positive from +25 to +50, Neutral from -25 to +25, Mostly Negative from -50 to -25, and Very Negative below -50. For how individual posts are classified as positive, negative, or neutral, see the sentiment analysis guide.

Sentiment distribution

The pie chart divides matching posts into positive, neutral, and negative; pointing at a segment shows its count and share. Unlike the score, the distribution includes neutral posts, which in many social media conversations form the largest block.

Read it to judge how much weight the score can carry. A strong score over a thin opinionated slice, visible as a large neutral segment, reflects few voices and can swing quickly; the same score over a broadly opinionated conversation reflects settled sentiment. The distribution also disambiguates a near-zero score, which can mean indifference (a dominant neutral segment) or a polarized standoff (large positive and negative segments balancing out). When the standoff pattern appears, the conversation is contested rather than calm, and the feed shows what each side is arguing.

The gauge shows which way opinion leans; the pie chart shows how much opinion there is.

Sentiment over time

The tab's centerpiece: how reception developed across the range, in three views.

• Score (the view the chart opens on) plots each interval's sentiment score as a line around a dashed zero line. The shape is the story of reception: slides, recoveries, and above all zero crossings, since the day the balance of opinion flips sides is a stronger signal than a small change in the average. It also puts the stat cells in context, showing whether the Most Negative Day was an isolated dip or the start of a slide. Intervals with little opinionated volume produce jumpy scores, so check suspicious swings against the Volume view before reading meaning into them.
• Volume stacks each interval's positive, neutral, and negative post counts. It explains score movement: a decline driven by new negative posts arriving is a different event than one driven by positive voices going quiet, and only this view tells them apart.
• Distribution normalizes each interval to 100%, isolating the opinion mix from the size of the conversation. A conversation that stays the same size while gradually turning negative is hard to spot in raw counts and obvious here.

Aggregate

Aggregate produces one reception verdict for the combined conversation and steadies the measurement along the way: variants that individually match too few opinionated posts to score reliably can produce a meaningful score together, so this is the better reading when the queries are spellings or phrasings of one subject.

Compare

The Compare view ranks the receptions and shows their trajectories side by side.

Stats bar

Most Positive and Most Negative name the queries with the highest and lowest sentiment scores, the two ends of the reception ranking. Because the score does not depend on volume, the ranking is fair between conversations of very different sizes; whether a lead is meaningful is something to confirm in the Sentiment chart below, where the Distribution view shows how much opinion stands behind each score.

Most Neutral and Least Neutral rank the conversations by the share of posts that take no side. This is the engagement axis the score misses: the least neutral conversation has the most opinionated audience, one that cares enough to judge, while the most neutral one is dominated by reporting and passing mentions, and its score rests on a thin opinionated slice. A subject can lose the score ranking while having the audience that feels most strongly, and these two cells are where that shows.

Sentiment

One bar per query, with a toggle between two views. Score plots each query's sentiment score on the -100 to +100 scale around a zero line; bars on opposite sides of the zero line mark genuinely opposed receptions rather than different degrees of the same one. Distribution stacks each query's bar to 100% with its shares of positive, neutral, and negative posts, showing how opinionated each conversation is. Reading the two together applies the same score-plus-distribution judgment as the single-query layout: a score lead built on a small opinionated minority is visible only in the second view.

Sentiment over time

One score line per query around the dashed zero line; this chart plots scores only. It puts the reception trajectories on one canvas, and the moments worth investigating are where the lines interact: a crossover marks the day one subject overtook the other, parallel dips suggest a shared shock such as industry-wide news, and how quickly each line recovers from one is a comparison of resilience that single-point scores cannot make. Narrow the time range to such a moment and use the Select view to read each side's posts.