Stats Chat

Team Ratings for Round 5

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 32 matches played, 19 of which were correctly predicted, a success rate of 59.4%.

Here are the predictions for last week’s games.

Predictions for Round 5

Here are the predictions for Round 5. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

Team Ratings for Round 8

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 47 matches played, 31 of which were correctly predicted, a success rate of 66%.

Here are the predictions for last week’s games.

Predictions for Round 8

Here are the predictions for Round 8. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

The Herald has pictures of the most beautiful faces in the world

and NPR reports on a computer algorithm that can tell if you sound trustworthy or calming or engaging.

The Herald story at least admits these faces are only world-famous in New Zealand (or, rather, the UK)

“It’s important to note that these are the idealised faces according to those living in the UK, so a study in Asia or Africa for example would no doubt have different results.”

The NPR story instead doubles down by saying

But algorithms have stamina, and they do not factor in things like age, race, gender or sexual orientation.

There’s a sense in which this is true, but it’s not a very useful sense. If we can guess age, race, or gender from the sound of someone’s voice, and these perceptions affect whether we think the voice is engaging,calming or trustworthy, our prejudices will show up in the training data and any competent black-box algorithm will learn them.

New South Wales has just voted, and the new electorate created where I lived in Sydney 20 years ago is being won by the Greens, who got 46.4% of the primary vote and currently 59.7% on preferences. The ABC News background about the electorate says

In 2-party preferred terms this is a safe Labor seat with a margin of 13.7%, but in a two-candidate contest would be a marginal Green seat versus Labor. The estimated first preference votes based on the 2011 election are Green 35.5%, Labor 30.4%, Liberal 21.0%, Independent 9.1, the estimated Green margin after preferences being 4.4% versus Labor.

There was definitely a change since 2011 in this area, so how did the polls do? Political polling is a bit harder with preferential voting when there are only two relevant parties, but much harder when there are more than two.

Well, the reason for mentioning this is a piece in the Australian saying that the swing to the Greens caught Labor by surprise because they’d used cheap polls for electorate-specific prediction

“We just can’t poll these places accurately at low cost,” a Labor strategist said. “It’s too hard. The figures skew towards older voters on landlines and miss younger voters who travel around and use mobile phones.”

The company blamed in the story is ReachTEL. They report that they had the most accurate overall results, but their published poll from 19 March for Newtown is definitely off a bit, giving the Greens 33.3% support.

(via Peter Green on Twitter)

The vertical axis on a bar chart must start at zero. The very rare exceptions are ones that prove the rule: where ‘zero’ isn’t zero. Otherwise, the axis starts at zero or it isn’t a bar chart. The whole point of bar charts is that the length of the bar is proportional to the data value.

Line charts and scatterplots are different.  They don’t need to be tied down to zero, and the axis scales can be chosen to make the information as clear as possible. With great power comes great responsibility, as we can see from the following pair of line graphs of oil drilling in the US.

It’s pretty obvious that these come from people with different communications agendas. Or, it would be, except they are from the same story at Bloomberg.

Neither graph has an ideal aspect ratio. The flat one is too flat: you can’t see the wobbles over time in number of rigs. The tall one is too tall: the number of rigs has halved, but it looks as though it has crashed much more than that.

Bill Cleveland has a useful default rule for scaling line graphs: the median slope of the line segments should be about 45 degrees. The orange line on the tall graph isn’t far off that, but the blue line is steeper.  The 45-degree rule would give a graph like this:

In fact, there is plenty of room to start the blue axis at zero, but that’s not always the right choice.

Here, in a sadly-appropriate pairing, is the Keeling Curve, the graph of atmospheric CO₂ concentrations at Mauna Loa observatory, in a visualisation paper from Berkeley.

There’s no sense at all in having the vertical axis start at zero. Zero is just not a relevant value of atmospheric CO₂. What’s more interesting, though, is how the two scalings show different information. The upper graph is scaled so the year-to-year changes have slope centred at 45 degrees. This makes it easier to see that the CO₂ increase is accelerating. The lower graph is scaled so the month to month changes have slope centred at 45 degrees, making it easier to see the shape of the seasonal pattern.

Different vertical scaling can be used just to mislead the reader, but it can also be used to make data more readable and to communicate more effectively.

• Two data-related notes about the Northland by-election: the polls were amazingly accurate given how hard by-elections are to predict, and the Electoral Commission did a wonderful job in getting the vote counted and reported fast.

• A graphic of gender inequality around the world.  The income graph is not the usual one, because it is total income, not income per hour worked — most of the gap is in fewer hours of paid employment.

• According to the Independent “California is in one of its worst-ever droughts because people are growing too much weed”. Unsurprisingly, not what the research actually says

• The Medical Council of New Zealand has released a Discussion Paper on the value of performance and outcome data.

• From Significance magazine, a description of how the various political polling companies in the UK work

Each week, we would like to invite readers of Stats Chat to submit nominations for our Stat of the Week competition and be in with the chance to win an iTunes voucher.

Here’s how it works:

• Anyone may add a comment on this post to nominate their Stat of the Week candidate before midday Friday April 3 2015.
• Statistics can be bad, exemplary or fascinating.
• The statistic must be in the NZ media during the period of March 28 – April 3 2015 inclusive.
• Quote the statistic, when and where it was published and tell us why it should be our Stat of the Week.

Next Monday at midday we’ll announce the winner of this week’s Stat of the Week competition, and start a new one.

(more…)

From the BBC, Hans Rosling on the Ebola epidemic

(That’s a diagram of the data collection system behind him)

(via Harkanwal Singh)

Stories were coming out recently about new cancer research led by Bryony Telford in Parry Guilford’s lab at Otago, and I’d thought I’d use it for an example of translation from Scientist to English. It’s a good example for news because it really is pretty impressive, because it involved a New Zealand family with familial cancer, and because the abstract of the research paper is well written — it’s just not written in ordinary English. Combining the abstract with the press release and a bit of Google makes a translation possible.

This will be long. (more…)

Q: Did you see that eating a bowl of quinoa every day helps you live longer?

A: No.

Q: There’s story on Stuff (well, from the West Island branches). Is it true?

A: Hard to say.

Q: Well, does the research claim it’s true?

A: Hard to say.

Q: Why? Didn’t they link?

A: No, they linked, and the paper is even open-access. It just doesn’t say anything about the effects of quinoa.

Q: But the story said “A new study by Harvard Public School of Health has found that eating a daily bowl of the protein-packed, gluten-free grain significantly reduces the risk of premature death from cancer, heart disease, respiratory disease and diabetes.”

A: Sadly, yes.

Q: This is your correlation and causation thing again, isn’t it?

A: No, the paper just doesn’t mention quinoa. It talks about grains and cereals.

Q: Ok. So they just didn’t break out the data for quinoa separately. It’s still a grain and a cereal, isn’t it?

A: Yes, as long as you aren’t even more pedantic than me. But it’s not just data analysis. They didn’t even ask their study participants about eating quinoa.

Q: So? Some of the grain they ate must have been quinoa, and there’s no reason to expect it’s different from other grains, is there? Won’t it all get averaged in somehow?

A: I suppose so. But there can’t have been that much of it getting “averaged in”

Q: Why not? You old folks may not have caught on, but quinoa’s getting popular now.

A: The study was in people over 50. That’s older than both of us. Even assuming we weren’t the same person.

Q: Even so. Things are changing. People have more adventurous diets. It’s not the twentieth century any more.

A: It is in the study.

Q: Huh?

A: The dietary data were collected in 1995 and 1997, from people with average age 61 years.

Q: Oh.