December 04, 2014
This follows on from this post, where I try to make some sense of my Pinboard bookmarks. In the last post I looked at monthly data but I wanted to get a bit more into it, so I reworked the Python script to get daily data. (I’m using the bookmarks file as a Python learning exercise so No, I don’t expect you to be impressed with my fledgling Python skills.)
from bs4 import BeautifulSoup
import datetime
bookmarks = BeautifulSoup(open('pinboard/bookmarks.html'))
counts = dict()
for link in bookmarks.find_all('a'):
epochdate = link.get('add_date')
epochdate = int(epochdate)
date = datetime.datetime.fromtimestamp(epochdate).strftime('%Y-%m-%d')
counts[date] = counts.get(date,0) + 1
def getkey(item):
return item[0]
results = sorted(counts.items(), key=getkey)
for day,count in results:
print day,count
I wrote that to a marks.txt
file and fed it into Gnuplot which gave me this:
Here’s the shell script I used for the plot.
#!/bin/sh
cat <<EOF | gnuplot
set terminal svg font "Helvetica,11"
set style data points
set xdata time
set grid
set yrange [0:]
set timefmt "%Y-%m-%d"
set format x "%Y-%m"
plot "marks.txt" using 1:2 t '' lw 3 lc rgb "#EF234D"
EOF
The uncommonly high (123) daily import was when I ran a script to import all of my GitHub starred repos into Pinboard so I tweaked the plot to ignore it, which gave me the kind of detail I was looking for.
I’m pleased to see that the overall trend is towards fewer big spikes in links saved and I would like to see that continue. I’ve become more ‘picky’ about the links I save and have gotten into a routine of regularly weeding, pruning and generally tending them. At least once a week I go through unread items and either read and archive or read and delete them, and make sure I have tags (that make sense) on anything I’m keeping. One tool that has made life easier is Ben Beckwith’s Pintaboard, a Pinboard plugin for Pentadactyl that makes bookmarking from Firefox, with tag completion, a breeze.