Contours should now be in roughly 1 minute increments. As discussed, this is assuming (erroneously, but what can you do?) that null values equal zero time spent looking for parking.
Comparing to the raw data (in QGIS), it seemed that this was generally underestimating times by quite a lot, so I reduced the bandwidth to compensate which, I perceive, got it a lot closer to estimated values where data was sparse. That may actually be the main noticeable change here, the bandwidth reduction.
Contours should now be in roughly 1 minute increments. As discussed, this is assuming (erroneously, but what can you do?) that null values equal zero time spent looking for parking.
Comparing to the raw data (in QGIS), it seemed that this was generally underestimating times by quite a lot, so I reduced the bandwidth to compensate which, I perceive, got it a lot closer to estimated values where data was sparse. That may actually be the main noticeable change here, the bandwidth reduction.