Systematic, Random, or Blunder? The Three Errors Hiding in Your Survey Data

All bad data looks the same at first glance: a number that's wrong. But professional surveyors don't treat every error the same way, because the fix for one won't touch the others. In a hands-on ROAR West workshop, surveyor and geodetic engineer Rami Tamimi, founder of The Survey School, broke down the three categories every data problem falls into — and why knowing which one you're looking at is the entire skill.

"Most of the time we can trace the problems in our data to three main error types," Tamimi said. The first is systematic error: a constant offset affecting the entire dataset in the same way, caused by something like a missing setting, a wrong reference point, or bad coordinates on a base station. Systematic errors are, in Tamimi's words, "the most undetectable, because everything looks fine, relatively speaking. It's not until you start to compare it to something else when that pops up." (watch at 7:45) The second is random error — small, unavoidable variation present in every measurement, which is why "there's no such thing as a perfect measurement." Random error can be quantified statistically, which is exactly what makes it manageable. The third is a blunder: a one-off human mistake, like transposing two digits or misreading a value, that produces an error large enough to stick out "like a sore thumb" once you know to look for it.

Tamimi's clearest explanation of the difference between accuracy and precision — a distinction that maps directly onto these three error types — came from a dad joke. "My kids asked me what's the difference between accuracy and precision. I told them accuracy is showing up to dinner on time, but precision is showing up 20 minutes late every single night." (watch at 8:55) Picture the classic dartboard example (watch at 9:50): tightly clustered darts that all land off to one side of the bullseye is precise but not accurate — systematic error, a constant shift. Darts scattered evenly around the bullseye, with the average landing dead center, is accurate but not precise — random error. A dart that lands nowhere near the others, with nothing to explain it, is a blunder.

Each type has a different fix. A systematic error, once identified, is a straightforward adjustment — Tamimi gave the example of an entire project shifted because a single base station had bad coordinates. (watch at 10:39) Because the error is constant, you can trace exactly where it came from and correct the whole dataset at once. Random error is managed statistically, using tolerance thresholds built on standard deviation (surveyors commonly reference one, two, or three sigma, corresponding to roughly 68%, 95%, and 99.7% confidence levels) (watch at 12:40). A blunder has no pattern to adjust for; the only defense is catching it, usually through a habit as simple as taking a measurement twice and checking that it makes sense. (watch at 15:10)

For anyone managing drone or GNSS survey data, the practical lesson is this: when something looks off, the first diagnostic question isn't "how do I fix this" — it's "which of the three is this." A dataset that's uniformly shifted points to a systematic cause worth tracing to its source. A dataset that's noisy but centers correctly points to normal random error within tolerance. A single wild outlier points to a blunder, and the fix is simply to find and remove it. Mistaking one for another wastes time chasing the wrong problem — or worse, applying a fix that corrects nothing at all.

Survey & mapping solutions · DJI drones for survey & mapping in Canada