Do people like, still blog?

Tag Archives: analytics

  • These Aren’t Wireless Headphones
    A look at Apple’s new wireless earbuds – if you’re like me and don’t really pay a lot of attention to Apple news, then you might know the ins and outs about this new product. Seems like it covers a lot of interesting use cases, but I’m not prepared to drop significant money on something that is so easy to lose (and only supports Apple devices)

    One more simple feature holds perhaps the most telling clue to what Apple has in mind for the future. Tap the AirPods twice while they’re in your ear and you’ll wake Siri, much like how you wake Amazon’s Echo by saying “Alexa.” Suddenly you’ll find yourself conversing with an A.I.–powered voice assistant via a tiny earpiece in your ear.

  • Hillary Clinton’s ‘Invisible Guiding Hand’
    It was surprising when I found out that analytics was such a big factor in a transient event such as an election, but after thinking about it, it makes a lot of sense. The data and analysis that has been accumulated can be reused for subsequent campaigns. However, I think it might be a waste that all the infrastructure might have to be recreated if the people are all new each time (I assume that there is a lot of custom analysis).

    The breakdown of the buy in Texas, powered by Kriegel’s modeling, shows how Clinton’s TV ads budget hunted for delegates, not votes. Texas is the rare state that used state legislative districts to award delegates, and Clinton spent $1.2 million on broadcast and cable ads even as she won the state by 32 percentage points. Sanders spent $0. She spent more on ads in tiny Brownsville ($127,000) and Waco ($142,000), ranked as the 86th and 87th largest media markets in the country, as she did in Houston ($105,000), the 10th largest, according to ad data provided by a media tracker.

    It paid off: In Texas alone, Clinton netted 72 delegates more than Sanders — a margin that more than offset all the Sanders’ primary and caucus wins through March 1.

  • Why Are Babies So Dumb If Humans Are So Smart?
    An interesting hypothesis as to why, when Humans are born, they’re so useless compared to other animals.

    And in modern humans, a few pieces of evidence appear to suggest that smarter parents are more likely to have offspring that survive. In one limited sample—two hundred and twenty-two Serbian Roma women—maternal I.Q. and child mortality were negatively correlated (that is, higher I.Q. meant lower mortality), even controlling for education, age, and a number of other factors. In a larger sample of Californian parents, in 1978, years of education were linked to infant-mortality rates. Global epidemiological studies suggest a decrease in mortality that equals between seven and nine per cent for each year of a mother’s education.

  • We might live in a computer program, but it may not matter
    I just blogged a similar article on this topic a few weeks ago, but this subject is so fascinating that I can’t get enough of it!

    Quantum mechanics, the theory of the very small, has thrown up all sorts of odd things. For instance, both matter and energy seem to be granular. What’s more, there are limits to the resolution with which we can observe the Universe, and if we try to study anything smaller, things just look “fuzzy”.

    Smoot says these perplexing features of quantum physics are just what we would expect in a simulation. They are like the pixellation of a screen when you look too closely.

  • The new science of cute
    Not surprisingly, this article is mostly about Japan – the epicentre of cute. There’s cute though, and there’s fame. This article tries to tackle both.

    But for a mascot to be successful, being cute is not always enough. For every popular yuru-kyara, there are a hundred Harajuku Miccolos – a 5ft-tall yellow-and-brown bee, who I met standing on the pavement outside the Colombin bakery and cafe, celebrating Honey Bee Day with three hours of loitering in front of the cafe, greeting passers-by, or trying to. Most barely glanced in his direction and did not break stride, though some did come over and pose for a photo. There was no queue.


One of the goals I set for myself for my Always Taeyeon app is to try and grow it better than A Healthier Commute. I already had a head start because the topic had more fans and audience, but I wanted the app to be “successful”. Of course, I set the normal hobby rules for the app (e.g., not paying for any promotion).

It’s been almost 2 months since I released and I’m pretty happy with the results so far. I’m at about 1300 downloads and a 4.4 star rating with 59 reviews. Obviously I want the download count to be higher, but I think I’m doing pretty well given that the only “promotion” I’ve done is post about it on a couple of Google+ and Facebook groups.

The statistics I’m most keen about now though is Daily Active Users. My app is high quality in the sense that the photos are high quality, but also in the sense that it is maintained well. So there are frequent (photo) updates. In fact, if there are new photos I try to engage with the user every day via a status bar notification. You can’t turn those notifications off – if you want them off, then you don’t care about the photos anyways so just uninstall the app. I’m not convinced that that is affecting my uninstall rate though as my install rate is a decent 42% (~570 current installs).

But does it have an affect on my DAU? I think it does. Here is a graph over time.

It is a nice trend as it is going up and to the right. I think my DAU is pretty impressive though as it is about 71% of my current installs! It’s so good that I’m kind of skeptical that the current install number is incorrect – although I haven’t released the app on other stores, maybe users are just passing around the APK.