I've created a walk-time map in ArcGIS Online, based on library locations within a city, using 5-minute increments between 0-5 mins and 25-30 mins. The attribute table shows these are stored as whole minutes in two columns (for each end of the range) e.g. 25.00 and 30.00 - this is what I expect.
I then want to summarise this data by finding the mean travel time (travel time band End value) within ~60 census areas of the city. I do this with Summarize Within.
The tool runs, but the output has values that are all less than 0 (range 0.06 to 0.68). Given that the lowest possible travel time value is 5, and the census areas are totally covered by travel time areas, how can this be?
There are also unexpected patterns in the data, where a census area that contains a library (has low walking time) shows with a (relatively) high value. See the lower-central library in the maps below.
I'm trying to write this up for a student exercise but simply cannot explain these results! Can anyone help?
Edit: I tried the same operation but summarising distance buffers instead of walk time. I have the same issue. Buffer distances are 500 1000 1500 2000 but the highest average in the summarised dataset is ~30 (should be just under 1500). And again that central area shows an odd, too-high result.