Widening Horizons — A year of reading
The start of a new year calls for some retrospective. And this time I decided to look back on how a whole year of my reading had fared and if I had managed to read sufficiently diverse topics over the course of the year.
While I continue to read off multiple sites, independent blogs, and books this year I happened to invest a large part of my blog reading time on Medium. When looking at the Medium stats on my profile the data available was surprisingly inadequate. It gives me view counters on stuff that I had interacted with or created and no real ability to look back at what I had been actually read during the course of the year or in the past.
And that is how this hack started. The objective of it being to derive insights on top of my reading history using which I could hopefully figure out what gaps I have had in my reading. To kick it off, I decided to define some objectives before I started out on the hack.
- What topics do I think I am interested in?
- How have my reading tastes modified or changed over time?
- Do I have any gaps in my reading?
- Any other interesting insights about my reading which can happen upon while I run the analysis.
I have been at heart interested in tech, entrepreneurship and engineering. This is what I do at my core, I also love travel (didn’t do much of it this year), love science fiction as a genre and like to read up and experiment with cooking once in a while. However, aside from all this I love to explore new ideas, topics and generally like a good well written story even if it ain’t fall into any of these buckets. Am not much politically inclined but satires always get me.
- Tech, Entrepreneurship,
- Machine Learning and Data Science
- Science Fiction
- Any Exploratory topics (surprise me here)
So first things first, getting the data in for analysis. Since I had done a bulk of my reading at Medium I decided to see if I could get my reading history from medium and run this post-analysis on top of it to figure how I had done overall on my interest areas.
The web-app didn’t have a history section and neither did the official medium api. However, I do a lot of my reading (almost all of it) on the mobile app and voila there was a section for the look-back, my very own reading history. It lacked an API but hell, that was a minor speed bump. Long story short I hooked up my device to go through a local proxy, and queried the medium history API. Voila! I had all my reading history with details of when I had read them and the article details themselves.
To set off I looked at some base stats from the reading history :
Data before May’s pretty sketchy but I think it may be owing to Medium launching it’s mobile android app during this time. December’s looking a bit of an outlier, but to be fair I was on 3 weeks leave, so no big personal surprises. I read more when I have the time to do so. I have held out at a steady 50 odd articles a month since the last 4 months so the work/personal schedule’s been fairly steady.
The average article length I have been reading has been on a downslide. Ain’t a good thing, this; points towards a decreasing attention span (gotta kick up that meditation and mindfulness schedule back to life). Something to tick off and watch more closely now over the coming days.
But the big Q remained, how had I done in terms of my reading diversity? Had i expanded my horizons and if so at all?
Now that I had the data, it was time to get to the underlying themes from the content I had read.
Medium’s tagging system allows each article to be attached with 3 topics. I have decided to rely on this classification and run my analysis on top of this. For now this was done using a manual topic to keyword lookup. At a later date probably I will derive topics using LDA or some form of automated topic extraction technique. Now that the aggregation was in place, it was time to derive topics from the keywords, e.g. Tech, startups, entrepreneurship, technology startups, Venture Capital all potentially imply a common theme around technology startups.
It was also time to look at the keywords and see if a natural taxonomy could be created which could them into broader topics/themes (especially around what I was interested in at the start.). Results as below. Probably need to work out a better visualisation for this data.
The stuff in green shows what have been far-out topics which I didn’t expect to read or touched/read upon for the first time.
The majority of content I read is around engineering and tech startups, no real surprises there. But quite a few surprises have cropped up. Themes like humour and love and marriage seem to be recurring. Didn’t really see that coming (the wife should be happy though). I seem to have completely or almost missed reading content around travel, cooking & sci-fi. Design seems to be a strong recurring theme (the inherent bias for well designed UX). I haven’t really touched on exploratory content or far out topics as much as I would like to.
While Medium has so far been a great place to discover new stories, yet I have missed good exploratory stuff to broaden my horizon in the news feed. The analysis of this reading history should hopefully allow me to tweak and re-engineer my library to fill up missing gaps. I hopefully know what I want to explore and read and often the lack of relevant content within a news feed makes it hard to pursue the love of reading the same. In the future, would love it if I could hook up my own little personal algorithm which I could tweak at will within my reading app. One which would help me recommend content based on around objectives like diversity, opinions, or just pure exploratory stuff around good stories. Probably an idea for a later hack.