The public map, which includes an e-mail notification service, was created as part of an overhaul of the LAPD's website funded with $362,000 raised by the Los Angeles Police Foundation.
The scale of the errors undermine the LAPD's expressed purpose in publishing the information. LAPD officials have said that a key reason for publishing the maps is so that residents can spot crime trends.
But earlier this year when the so-called Salt & Pepper Bandit robbed 13 Los Angeles banks in less than a month, only six of the heists found their way onto the LAPD map.
In Atwater Village, Insalaco and other residents saw that about half the crimes reported by the news media had failed to appear on the LAPD's site. But when their concerns about a "massive misinformation campaign" came to the attention of local police officials, Northeast Division Capt. William A. Murphy dismissed them.
"The simple answer -- no," Murphy wrote in an April 22 newsletter responding to Insalaco's allegations. "There is no way that we under report crimes in [the Northeast division] by 50%."
Contacted Wednesday, Murphy said he was troubled by the map's flaws.
"From my perspective, any time it's even slightly off, that can cause a question on all the statistics," Murphy said.
He conceded, though, that he did not check the map for accuracy even though he urged community groups to use it.
"I'm a police officer, I'm relying on the technology," Murphy said. "I'm not an expert on computer systems."
Those who are experts on technology say data omissions on this scale suggest a lack of attention to accuracy.
It means "someone was not being particularly concerned about their data quality," said Clay Johnson, director of software development for the nonprofit Sunlight Foundation, which republishes government data online and urges the government to release more.
Still, Johnson commends government agencies that make an effort to post data online.
Although he called a 40% data loss a big problem, he added that some level of error is unavoidable when working with complicated databases.
"I think it is acceptable as long as it's correctable," Johnson said. "You have to accept failure because it's going happen."
--
ben.welsh@latimes.com
doug.smith@latimes.com
