I think the redesign of the output from sheets may be the culprit here. The time trend for China in this sheet is the opposite of the data, but the column output for the year 1997 seems to grow as you'd expect China's patents to rise over time.
Here's another example. I aggregated data by sorting and deleting in Excel, so the CSV has the actual numbers that can be compared to what Ravel is pumping out. There are obvious cases of numbers being scambled.
Good catch. When I transposed the sheet so that it aligned with the ravel handles, I basically just swapped the indexes 0 & 1. Unfortunately, in C++, you don't have multidimensional arrays (they're coming in C++23 I believe), so you have to calculate a linear index of the form i+j*dim[0]. I swapped the 0 for a 1 instead of swapping the i and j. It took an egregious dataset such as this patent dataset for that mistake to be obvious.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
My gawd, you're human after all! 😉. That sounds like one of my mistakes!
Good that we've caught this one now, anyway. Can you get a new beta to me
ASAP? I want to use this dataset as an indication of getting insights
rapidly from Ravel on the Patreon site.
CSV file again.
More of the same: sheet and plot returning different numbers when displaying the same data.
I think the redesign of the output from sheets may be the culprit here. The time trend for China in this sheet is the opposite of the data, but the column output for the year 1997 seems to grow as you'd expect China's patents to rise over time.
Here's another example. I aggregated data by sorting and deleting in Excel, so the CSV has the actual numbers that can be compared to what Ravel is pumping out. There are obvious cases of numbers being scambled.
RVL and CSV
Good catch. When I transposed the sheet so that it aligned with the ravel handles, I basically just swapped the indexes 0 & 1. Unfortunately, in C++, you don't have multidimensional arrays (they're coming in C++23 I believe), so you have to calculate a linear index of the form
i+j*dim[0]
. I swapped the 0 for a 1 instead of swapping the i and j. It took an egregious dataset such as this patent dataset for that mistake to be obvious.My gawd, you're human after all! 😉. That sounds like one of my mistakes!
Good that we've caught this one now, anyway. Can you get a new beta to me
ASAP? I want to use this dataset as an indication of getting insights
rapidly from Ravel on the Patreon site.