Posts filed under Social Media (95)

From Stuff,

If you’ve ever viewed Twitter as a gauge of public opinion, a weathervane marking the mood of the masses, you are very much mistaken.

That is the rather surprising finding of a new US study, which suggests the microblog zeitgeist differs markedly from mainstream public opinion.

Apart from being completely unsurprising, this is a useful thing to have data on.  The Pew Charitable Trusts, who do a lot of surveys, compared actual opinion polls to tweet summaries for some major political and social issues in the US, and found they didn’t agree.

Along the same lines, it was reported last month that Google’s Flu Trends overestimated the number of flu cases this year (after having initially underestimated the H1N1 pandemic), probably because the high level of publicity for the flu vaccine this year made people more aware.

These data summaries can be very useful, because they are much less expensive and give much more detail in space and time than traditional data collection, but they are also sensitive to changes in online behaviour. Getting anything accurate out of them requires calibration to ‘ground truth’, as a previous generation of Big Data systems called it.

Juha Saarinen, on Twitter, asks for data addresses the Auckland rent shock headlines.

Detailed data on new rentals are available from the ‘market rent’ pages at MoBIE, but there doesn’t seem to be any way to download the whole thing, just individual neighbourhoods.  And it’s a long weekend.

The Herald has a story about a new app called TalkTo. Rather than you calling a business and waiting around for a possibly unhelpful response, you can text TalkTo and wait for them to call the business, ask your question and pass on the unhelpful response. Or, at least, you can if the business is in the USA or Canada — they currently wouldn’t handle Novapay or Qantas, the two examples in the story. The app obviously wouldn’t help for issues that require a dialogue, which includes essentially all the time I spend on hold.

Anyway, the statistics angle is that we apparently spend 43 days on hold during our lives.  As a basic numeracy challenge: is this more than you expect or less?

The number comes from 20 minutes per week for 60 years, so it doesn’t apply to any actually existing people — 60 years ago, we didn’t have the same level of on-hold, and 60 years in the future there’s at least some hope that a larger fraction of businesses will figure out how to make a useful web page (or whatever the next communication technology but seven turns out to be).

Nate Silver is doing a Q &A session at reddit.com (they have a running feature I am A …. ask me anything)

Stuff says

Facebook’s Instagram lost almost a quarter of its daily users a week after it rolled out and then withdrew policy changes that incensed users who feared the photo-sharing service would use their pictures without compensation.

There’s two things wrong with this (apart from almost-explicit post hoc ergo propter hoc).  The first is that the ‘quarter of its daily users’ is actually an estimate based on people who use Instagram in a way that shows up on AppData’s counters for Facebook apps. As AppData says

This application is integrated into Facebook from one or more platforms outside the Facebook.com canvas. As such, only users who connect to the app using Facebook are included in the active user counts above & below.  

The Stuff story actually admits to this later, contradicting the lead. Even with the data just coming from a non-representative sample, the change in use is pretty dramatic, but the phrase ‘a week after’ in the story is also important.  AppData shows just two weeks of data free, but we can put together the Dec 14-28 graph shown by Quartz when they covered this with the current graph from Dec 17 to now:

The new license agreement came out on Dec 17.  Nothing happened for a week, then there was a decrease.  On December 29 there was another decrease.

AppData’s results for weekly active users of Instagram didn’t change much over this period, and other apps also saw a decrease in daily users via Facebook — Stuff mentions Yelp, but I also saw it for Scribd, Spotify, Bing, and TripAdvisor.  In fact, StatsChat has also seen a decrease in users via Facebook over the past week.

It could be that Instagram’s license mistake is reponsible for its decrease, but at this time of year there are other possible explanations for people changing their computer use habits. We’ll be able to tell in a month or so whether the decrease is persistent. Perhaps Stuff can revisit the issue then.

Andrew Gelman has a review of two reviews of Nate Silver’s book, The Signal and the Noise. Unlike him, I’ve actually read the book, but I think his review of the reviews captures the good and bad points well.

The graph below is an overly-creative variation on barplots, which I think confirms the principle “if you want to write the data values on the graph, it’s probably a bad graph”.

The data are supposed to be “time per day spent using mobile apps”.  Presumably it’s mean time per day, though I can’t tell whether the mean is restricted to people who spend non-zero time.  The graphs come from a “study” conducted by the “Science Advisory Board®”.  The “Science Advisory Board®” is an online survey panel for market research, where biomedical scientists are the market.  Or as they put it

The Board is an independent, worldwide panel of life science and medical professionals that convenes electronically to voice their opinions on a wide range of topics.

Here “convenes electronically” means “gets sent survey links by email”, and since I’m not a “member”, in my case this means spam about a survey on lab equipment.

The “Science Advisory Board®” homepage includes various things aimed at making their study samples feel like a community. There’s also widget that cycles through their past few days of Twitter feed at a rate of about one every five seconds, and, rather surprisingly for a company that wants to give the impression of solid opinion research, a clicky bogus poll™.

Via Stuff, Twitter heat maps composed by SGI, showing positive and negative sentiment on Twitter on particular topics.

This one is from the US election, and it shows the good and bad aspects of the heatmap.  Since the information is in the colour scale, you don’t have the problem we saw earlier this week

On the other hand, you do have the problem that high population density regions are the ones that show up — giving a perhaps-misleading impression in this image that there was overwhelmingly more positive sentiment than negative about the US election results.

[update: wrong map initially]

XKCD on the data visualisation equivalent of forgetting that Auckland is larger than Wellington

I’m looking at you, Facebook

A Herald story today tells us, four times,  “22 per cent of women feel they have stronger relationship with girlfriends than their partners.”  According to a 2009 survey (PDF), 18% of adult Kiwis were not married, living together, or even dating.  That could potentially explain a lot of the 22%.

At least that one was for a charitable cause. The other survey story was for an online dating service that relies on recommendations from (Facebook) friends. Unsurprisingly, the survey they commissioned could be read as supporting this strategy:

43 per cent of people trusted their friends and families when it comes to dating advice, tips and recommendations.

and

The survey found singles were less inclined to go to a professional matchmaker for help, with only 1 per cent saying they would trust them