From the perspective of the Erik Wemple Blog, Donald Trump’s campaign has consisted of one scandalito after another, relating to his many, many falsehoods, bigoted statements and unhinged Twitter outbursts.

The analytical machine at research firm mediaQuant is seeing something altogether different:

See that green line? It indicates that for the month of April, Trump’s media coverage was 85 percent positive — a number that Paul Senatori, mediaQuant’s chief analytics officer, says has been pretty consistent throughout Trump’s nearly year-long campaign.

Now for the corresponding figure for Hillary Clinton:

“It’s amazing that Hillary Clinton has gotten more negative sentiment than Trump has gotten,” says Senatori.

Just what articles and television reports drive those ratings? Quite a few. MediaQuant looks at 14 “traditional and social media segments, e.g., worldwide newspapers, TV broadcast, blogs, Twitter, consumer media, etc., culled from more than 100,000 global news sources.” And how does it digest all that information and determine whether it’s positive, negative or neutral? Via machine: Positive-negative calls, notes the company, are “based upon the prominence (volume + relative importance) of industry-standard ‘positive’ or ‘negative’ phrases and their proximity to the topic mention.” Media-analysis nerds call the methodology “natural-language processing.”

Senatori provided this example of how the coding system would detect a negative story: “If there is language … that talks about fraud or indictments or lying or misrepresenting,” the mediaQuant system might well flag the piece of content as “leaning toward negative.” It’s not hard to think about numerous news stories that deploy such terms in coverage of both Clinton and Trump. In any case, we asked Senatori if it’d be possible to run a specific article through the mediaQuant maw — like, for example, last week’s Post article about Trump masquerading as his own flack. The program doesn’t accommodate such one-offs, said Senatori.

Folks unfamiliar with name “mediaQuant” may be familiar with its work, at least in this election cycle. Its computer systems birthed the figure that Trump had earned $2 billion in what some people called “free media” as of late winter. The New York Times wrote up the figure, and so did the Huffington Post and CNN. And Ted Cruz: “With all due respect the media has given $2 billion of free advertising to Donald,” Cruz said in an appearance on “Meet the Press.

MediaQuant prefers the term “earned media” or “media attention.” And to update the numbers, Trump’s “media attention” tally through the end of April stands at $2.8 billion. Clinton clocks in at $1.1 billion, and Bernie Sanders at $658,000.

The Erik Wemple Blog has cast a skeptical eye at Trump’s blowout lead on this front, as calculated by mediaQuant. We’ve sampled a chunk of Trump coverage and noted it is often skeptical of his claims, his trustworthiness or his qualifications for the presidency. Is it possible that negative coverage of Donald Trump has padded his “free media” bounty? MediaQuant has a means of guarding against that prospect, notes Santori. Any “negative” story about Trump (or any other candidate) counts for about one-third the value of a “positive” story about Trump in calculating his overall earned media total. Indeed, mediaQuant has counted more than 273,000,000 Donald Trump “mentions” over the past year, and if the positive-negative breakdown of April has been consistent with previous months (as Senatori says it has been), that translates into about 41 million negative “mentions.” The firm has tracked 90 million total “mentions” for Clinton over the same period, and if the April breakdown is consistent with previous months, she has racked up about 27 million negative “mentions.”

So: All press is good press, at least on some fractional level.