Posts filed under General (515)

August 1, 2014

The algorithm of life

“…Unlike people, computer programs aren’t embarrassed by their prejudices and don’t try to hide them.” Thomas Lumley discusses the algorithm-based tools that sift data about us and our behaviour in this week’s New Zealand Listener. Click here for the full story.

 

July 31, 2014

Briefly

‘This is statistics’ website

The American Statistical Association is launching a public relations campaign to make people think statistics is less boring and pointless, which is good:

We want students and parents to have a better understanding of a field that is often unknown or misunderstood. Statistics is not just a collection of numbers or formulas. It’s not just lines, bars or points on a graph. It’s not just computing. Statistics is so much more. It’s an exciting—even fun—way of looking at the world and gaining insights through a scientific approach that rewards creative thinking.

That’s a quote from the  shiny new website, ThisIsStatistics. It has stories about what statisticians do, and information about salary and job trends and stuff.  There are videos of statisticians talking about their work: currently Roger Peng (Johns Hopkins, SimplyStatistics blog) and Genevra Allen (Rice University).

It’s slightly disappointing that more of the people on the site arent’ real, just stock photos, but I suppose that’s unavoidable. What’s a bit more annoying is one of the photos in particular:

About-ASA-cropped

This looks as if it was constructed specially (the cup/mat/tablet/glasses are stock, eg).  It’s a rose chart, which is an ok way to display circular data (eg wind directions), but is not so good for comparison because of the way the wedges change shape as they get larger. The numeric labels are also a slightly strange choice for a circle measured in degrees (90 isn’t a multiple of 20).

Much more importantly, given the emphasis of the site on statistics as solving real problems, this is labelled as not being real: “data A” and “data B”.  Not helpful when we’re trying to tell people “It’s not just lines, bars or points on a graph”.

 

July 30, 2014

NRL Predictions for Round 21

Team Ratings for Round 21

The basic method is described on my Department home page. I have made some changes to the methodology this year, including shrinking the ratings between seasons.

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

Current Rating Rating at Season Start Difference
Sea Eagles 9.18 9.10 0.10
Rabbitohs 7.48 5.82 1.70
Roosters 7.09 12.35 -5.30
Cowboys 4.89 6.01 -1.10
Warriors 4.38 -0.72 5.10
Storm 2.94 7.64 -4.70
Broncos 0.81 -4.69 5.50
Panthers -0.02 -2.48 2.50
Bulldogs -0.76 2.46 -3.20
Knights -1.30 5.23 -6.50
Dragons -2.44 -7.57 5.10
Titans -5.06 1.45 -6.50
Raiders -6.23 -8.99 2.80
Wests Tigers -6.81 -11.26 4.50
Sharks -7.26 2.32 -9.60
Eels -8.66 -18.45 9.80

 

Performance So Far

So far there have been 144 matches played, 79 of which were correctly predicted, a success rate of 54.9%.

Here are the predictions for last week’s games.

Game Date Score Prediction Correct
1 Knights vs. Roosters Jul 25 16 – 12 -5.80 FALSE
2 Broncos vs. Storm Jul 25 8 – 30 7.40 FALSE
3 Panthers vs. Sharks Jul 26 16 – 18 14.80 FALSE
4 Titans vs. Eels Jul 26 18 – 24 11.20 FALSE
5 Bulldogs vs. Cowboys Jul 26 12 – 20 0.50 FALSE
6 Warriors vs. Sea Eagles Jul 27 12 – 22 1.90 FALSE
7 Wests Tigers vs. Dragons Jul 27 12 – 28 3.60 FALSE
8 Raiders vs. Rabbitohs Jul 28 18 – 34 -7.60 TRUE

 

Predictions for Round 21

Here are the predictions for Round 21. 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.

Game Date Winner Prediction
1 Sea Eagles vs. Broncos Aug 01 Sea Eagles 12.90
2 Bulldogs vs. Panthers Aug 01 Bulldogs 3.80
3 Sharks vs. Eels Aug 02 Sharks 5.90
4 Cowboys vs. Titans Aug 02 Cowboys 14.50
5 Roosters vs. Dragons Aug 02 Roosters 14.00
6 Raiders vs. Warriors Aug 03 Warriors -6.10
7 Rabbitohs vs. Knights Aug 03 Rabbitohs 13.30
8 Wests Tigers vs. Storm Aug 04 Storm -5.20

 

Super 15 Predictions for the Super Rugby Final

Team Ratings for the Super Rugby Final

The basic method is described on my Department home page. I have made some changes to the methodology this year, including shrinking the ratings between seasons.

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

Current Rating Rating at Season Start Difference
Waratahs 10.21 1.67 8.50
Crusaders 10.20 8.80 1.40
Sharks 3.91 4.57 -0.70
Hurricanes 2.89 -1.44 4.30
Bulls 2.88 4.87 -2.00
Chiefs 2.23 4.38 -2.10
Brumbies 2.20 4.12 -1.90
Stormers 1.68 4.38 -2.70
Blues 1.44 -1.92 3.40
Highlanders -2.54 -4.48 1.90
Lions -3.39 -6.93 3.50
Force -4.67 -5.37 0.70
Reds -4.98 0.58 -5.60
Cheetahs -5.55 0.12 -5.70
Rebels -9.53 -6.36 -3.20

 

Performance So Far

So far there have been 124 matches played, 82 of which were correctly predicted, a success rate of 66.1%.

Here are the predictions for last week’s games.

Game Date Score Prediction Correct
1 Crusaders vs. Sharks Jul 26 38 – 6 7.40 TRUE
2 Waratahs vs. Brumbies Jul 26 26 – 8 9.40 TRUE

 

Predictions for the Super Rugby Final

Here are the predictions for the Super Rugby Final. 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.

Game Date Winner Prediction
1 Waratahs vs. Crusaders Aug 02 Waratahs 4.00

 

July 29, 2014

H.G. Wells and statistical thinking

July 28, 2014

The Games: How we’re doing

Statistics New Zealand is running the numbers during the Glasgow 2014 Commonwealth Games to show how many medals countries are winning relative to their population.  At the time of posting, we were third on a per-million-of-population basis. Check it out here.

NZ Data Futures Forum: Discussion paper out

The New Zealand Data Futures Forum, which was established by the Ministers of Finance and Statistics to explore the future of data-sharing between the public and private sector, has released a discussion paper here.

This is the press release that was issued this morning:

Paddock to plate, and smart roads possible – NZ Data Futures Forum

New Zealand’s international brand and exports could grow significantly with the creation of a data sharing ‘eco-system’ according to a paper released by the NZ Data Futures Forum today.

Food traceability or ‘paddock to plate’ tracking is one of a number of kick start projects recommended in the paper that would see New Zealand become a world leader in the trusted use of data.

“New Zealand has got a real opportunity here. If we can create an ‘eco-system’ for data, we can unlock huge value, but to do this we need to treat data as a national asset,” says Forum Chair John Whitehead.

The paper suggests a range of initiatives including the establishment of an independent data council and an open data champion to drive innovation through data sharing.  The data council would act as an independent ‘guardian’ to ensure trust, privacy and security are maintained.

“Getting the rules of the game right is a vital part of encouraging collaboration, creativity and innovation.  New Zealand is uniquely placed to do this extremely well.”

The development of ‘smart roads’ that pull data from a range of sources, such as cats eye data capturing traffic flow, is another example the Forum uses to highlight the value that can  be created through collaborative data sharing.

“Transport is a critical issue for Auckland. Smart roads can keep traffic moving more freely and prevent a future of bottlenecks and delays literally putting a brake on productivity

“If our recommendations are followed we will see New Zealand lead the world in this space. The potential gains are limitless, including the ability to tackle immediate and real social problems.”

 

July 23, 2014

NRL Predictions for Round 20

Team Ratings for Round 20

The basic method is described on my Department home page. I have made some changes to the methodology this year, including shrinking the ratings between seasons.

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

Current Rating Rating at Season Start Difference
Sea Eagles 8.06 9.10 -1.00
Roosters 8.02 12.35 -4.30
Rabbitohs 6.67 5.82 0.90
Warriors 5.50 -0.72 6.20
Cowboys 4.07 6.01 -1.90
Broncos 3.33 -4.69 8.00
Panthers 1.50 -2.48 4.00
Storm 0.42 7.64 -7.20
Bulldogs 0.06 2.46 -2.40
Knights -2.24 5.23 -7.50
Titans -3.51 1.45 -5.00
Dragons -4.19 -7.57 3.40
Wests Tigers -5.06 -11.26 6.20
Raiders -5.42 -8.99 3.60
Sharks -8.78 2.32 -11.10
Eels -10.21 -18.45 8.20

 

Performance So Far

So far there have been 136 matches played, 78 of which were correctly predicted, a success rate of 57.4%.

Here are the predictions for last week’s games.

Game Date Score Prediction Correct
1 Sharks vs. Cowboys Jul 18 18 – 36 -6.10 TRUE
2 Eels vs. Rabbitohs Jul 18 12 – 32 -10.60 TRUE
3 Roosters vs. Panthers Jul 19 32 – 12 8.90 TRUE
4 Storm vs. Raiders Jul 19 28 – 14 9.40 TRUE
5 Broncos vs. Warriors Jul 19 28 – 22 1.40 TRUE
6 Knights vs. Titans Jul 20 8 – 22 10.00 FALSE
7 Wests Tigers vs. Bulldogs Jul 20 46 – 18 -6.40 FALSE
8 Dragons vs. Sea Eagles Jul 21 12 – 21 -7.40 TRUE

 

Predictions for Round 20

Here are the predictions for Round 20. 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.

Game Date Winner Prediction
1 Knights vs. Roosters Jul 25 Roosters -5.80
2 Broncos vs. Storm Jul 25 Broncos 7.40
3 Panthers vs. Sharks Jul 26 Panthers 14.80
4 Titans vs. Eels Jul 26 Titans 11.20
5 Bulldogs vs. Cowboys Jul 26 Bulldogs 0.50
6 Warriors vs. Sea Eagles Jul 27 Warriors 1.90
7 Wests Tigers vs. Dragons Jul 27 Wests Tigers 3.60
8 Raiders vs. Rabbitohs Jul 28 Rabbitohs -7.60

 

Super 15 Predictions for the Semi-finals

Team Ratings for the Semi-finals

The basic method is described on my Department home page. I have made some changes to the methodology this year, including shrinking the ratings between seasons.

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

Current Rating Rating at Season Start Difference
Waratahs 9.66 1.67 8.00
Crusaders 8.78 8.80 -0.00
Sharks 5.34 4.57 0.80
Hurricanes 2.89 -1.44 4.30
Bulls 2.88 4.87 -2.00
Brumbies 2.76 4.12 -1.40
Chiefs 2.23 4.38 -2.10
Stormers 1.68 4.38 -2.70
Blues 1.44 -1.92 3.40
Highlanders -2.54 -4.48 1.90
Lions -3.39 -6.93 3.50
Force -4.67 -5.37 0.70
Reds -4.98 0.58 -5.60
Cheetahs -5.55 0.12 -5.70
Rebels -9.53 -6.36 -3.20

 

Performance So Far

So far there have been 122 matches played, 80 of which were correctly predicted, a success rate of 65.6%.

Here are the predictions for last week’s games.

Game Date Score Prediction Correct
1 Brumbies vs. Chiefs Jul 19 32 – 30 5.00 TRUE
2 Sharks vs. Highlanders Jul 19 31 – 27 13.00 TRUE

 

Predictions for the Semi-finals

Here are the predictions for the Semi-finals. 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.

Game Date Winner Prediction
1 Crusaders vs. Sharks Jul 26 Crusaders 7.40
2 Waratahs vs. Brumbies Jul 26 Waratahs 9.40