ENOSUCHBLOG

Programming, philosophy, pedaling.

• Home
• Tags
• Series
• Favorites
• Archive
• Main Site

A Month of Robocalls

Oct 3, 2018     Tags: curiosity, data    

I’ve heard a lot of people remark (both online and offline) about the volume of robocalls they’ve received lately¹, including ones that fake either the area code or central office code to make the call appear as if it has a local origin. I also felt that I was receiving more robocalls (and more robocalls with fake numbers) than I had before, although I had no data to back that up.

So, I kept a log of my spam calls for a month.

Caveats

• This obviously isn’t scientific — it’s just for fun.
• I’m not always in areas with cell service. While I picked up every call that made it to my phone in the past month, I did not attempt to go through my voicemail for robocalls.

The data

Everything below is in Eastern Standard Time (EST).

Since these numbers are probably spoofs of actual numbers, I’ve redacted the last digit of each office code and two digits of each line:

Overview

Between August 29th and October 3rd (36 days, so about 1.25 months), I received 31 robocalls, or about 0.86 robocalls a day.

However, I received exactly 0 robocalls on weekends. Counting only the 26 weekdays, I received an average of 1.2 robocalls a day, or about 6 robocalls per work week.

Hours

The earliest call occurred on 9/12, at 9:16 AM. The latest occurred on 9/18, at 5:56 PM. Unsurprisingly, all calls occurred within business hours plus one (9AM to 6PM).

The average time for all calls was 13:09.

For morning calls (n=12), it was 10:39; for afternoon (n=19), 14:43.

Area and office codes

The most common area code (by a factor of two!) was 201 (allocated for all of New Jersey, but most common in the Hudson/Bergen area right across the river from NYC). Interestingly, the majority of the 201 calls (69%) occurred on the week of 9/11 and shared the same office code (20<REDACTED>) — my guess is that this particular robocaller latched onto the fact that I was picking up.

The next most common was 917, which is one of Manhattan’s four area codes (the others are 212 (n=1), 646 (n=0), and 332 (n=0)). This one makes sense to me: it’s my area code, and so a robocall operator has a good chance of hitting a number that at least looks familiar to me.

The rest of the area codes occurred only once, and are all either upstate NY, the tristate area, or off-by-ones: 681 (West Virginia) is one-off from 680 (which shares the same region as 315 (n=1)), and 918 (Oklahoma) is one-off from 917 (Manhattan, n=8). This strikes me as a strange strategy: area codes are much more distinctive than office codes or line numbers, so changing them at all makes people notice that there’s something weird about a call.

Content

Unfortunately, I neglected to record the calls. However, from memory:

• A good chunk of calls, maybe the majority, were related to health insurance: opportunities to lower my premiums, increase my coverage, and so forth. I’m guessing that this is spurred on by open enrollment periods (many private insurers have just finished theirs, and the federal government’s begins on November 1).

• Another good chunk began with a command: “Do not hang up.” I think I hung up on all of those.

• Finally, a sizable minority were in what sounded like Mandarin (to my untrained ear). I don’t speak Mandarin, so I have no idea what they were saying. WNYC has reported on robocalls that sound similar, as part of an extortion/blackmail scam.

Wrap-up

31 calls isn’t all that great of a sample size, but there were still plenty of interesting patterns: robocallers apparently have enough success with NJ area codes in NY to continue trying them, and at least one of them mistook my (muted) pickups for an engagement and continued to call me with the same area and office codes for a whole week. There isn’t enough data to say that robocallers prefer making afternoon calls, but I wouldn’t be surprised if the current pattern held up.

I also kept track of spam texts for the month, but the sample size (n=5) was way too small to infer anything. They also stopped after 9/2, which makes me think that I was only part of a relatively small spam campaign.