When you realize data.table is not just fast…
but also lets you:
- filter like a ninja
- group like a boss
- update in place like magic
- chain like you're riding 🚲 downhill
And all with [i, j, by].
Name a tidier feeling. I’ll wait.
#rstats#datatable
I took a photo of a USAID nutrition pack. Nothing weird… until I looked closer.
The image had hidden GPS data.
Using Python, I pulled the coordinates and put a pin on the map.
What I found was… strange.
Should foreign aid be ending up here? 👇
ALT An USAID nutrition pack in the right place? Is this evidence of fraud?
I mapped the exact spot where the USAID nutrition pack was photographed.
You’ll see the coordinates, the map, and how I did it in Python.
This isn’t just about one photo — it’s about transparency.
📍 Full post & map here →
I found a USAID nutrition pack lying around and took a photo — just out of curiosity.
Later, I checked the image’s metadata…
It turns out that GPS coordinates were embedded in it. 👀
So I mapped it. See below for the location! 1/3
I broke down how to extract location data from any image — and revealed the exact location of this USAID pack.
🧭 Want to see where it was taken?
📍Check the map story here:
marsja.se/how-to-extract-gps…
Should it be there? Let me know!
3/3
Filtering large datasets in R? You might be using dplyr::filter(), but there's a faster way. 🚀 Let's compare dplyr and data.table for big data filtering. 👇 #rstats#DataScience
To find out, we ran a benchmark comparing dplyr::filter() vs data.table on a 20-million-row dataset.
We filtered rows where a > 0.5 and recorded the execution time. 🚀
Here's what happened...