Recently I have switched my research towards the relationship between agriculture and greenhouse gas emissions. We all need to eat and climate change is our biggest problem, so why wouldn’t I? Malcolm and I are working on the effects of reducing the amount of tillage. Here is a map using Statistics Canada data of zero till.
Eurostat has published the year on year changes in house prices in European countries. I have used Tableau to visualize them. The darker the green, the greater the positive change. Red or pink means a drop (so sorry La France). If you want a map over which you can hover your cursor and get the actual change, go here https://public.tableausoftware.com/shared/B2H6QP5SJ?:display_count=no
Data for some countries (Germany and Poland and Greece) is missing, which is why there are those big holes in Europe.
House price percentage increases 2013 to 2014
Just started learning how use Tableau for mapping. First effort uses World Bank figures on train use and car ownership. Click on the map tab top right. Use the sliders to the right of the map to change the parameters. Darker green shows more millions of passengers per km of railway. You can filter by ownership of passenger cars per 1000 population.
https://public.tableausoftware.com/views/Transportation_4/Sheet2?:embed=y&:display_count=no
Just had an article published which concerns the prediction of just which students are going to take up their offers of a university place. Registrars spend $$$ sending out offer letters to students who don’t show up. I use GIS and stats to build a classification tree model. You can feed in a whole bunch of student data, and then get a probability of offer takeup. Then you can decide who to target with scholarships etc. To get this article, go to the publications tab and scroll down to peer-reviewed journals. You can download the pdf for free there.
The 1836 Tithe Commutation Commission files contain a huge amount of highly useful data, including the pastoral rents and yields of some parishes in the south-west of England. I want to see how the rent changed with distance from the main market, which was London. No trucks or railways then, so the animals had to be driven overland. This was expensive no least because the animals was sold by weight and a long journey caused significant weight loss. The regression equation works fine. I used elevation as an independent variable, and the full equation was: Pasture Rent = elevation + Pasture Yield + London Distance. As expected, elevation and London Distance had negative signs. A map of the parishes colored with their rent is below. Only three categories shown. The darker the blue the higher the rent. The escarpments that existed at that time are also shown.
. Notice how the higher rents are all closer to London. But how about the non-linear effects of elevation and distance? If the farm was very high up then the animals would have to go up hill and down dale? I used the mgcv package in R for this. Here is a contour plot showing the rents:
Look at the contour lines where you can see the rents. They are curved especially about halfway along. This reflects the topography of Devon.
The yields of crops such as wheat are clearly affected by the amount of moisture available in the soil, as well as the temperature. More if generally better, but up to a point. Yields increase and then drop off. It looks like a reversed U-shaped curve. The usual way of modeling this would be with a quadratic term, that is including a squared term in the regression. That has been the approach taken by leading reasearchers looking into the effects of climate change on agriculture. But what if the two variables act on each other in a non-linear fashion? I have been working on this problem using regression splines, allowing the relationship to be flexible or ‘wiggly’. The estimation is non-parametric but so far seems to show more accurate results than the rather rigid parameterization of the quadratic method. The plot below shows wheat yields from the 1830s, against temperature and precipitation in August. The non-linearity is clear. I used the mgcv package in R for this. 
Trying to predict stockmarket prices over time is a bit like the old alchemists. Being told that you can’t do it just makes you try harder. Here is a plot of the NYSE in 1996. The code etc is on my RPubs site, herehttp://rpubs.com/Stephen_P/10377
I just found the quantmod package which retrieves share prices and volatility with just a few lines of code and then makes a beautiful plot. Here is Yahoo with Bollinger Brands (Click on the link if the plot doesn’t appear): 
<a href=”http://geostatskpu.files.wordpress.com/2013/09/yahooshares.png”><img src=”http://geostatskpu.files.wordpress.com/2013/09/yahooshares.png?w=300″ alt=”yahooshares” width=”300″ height=”130″ class=”alignnone size-medium wp-image-134″ /></a>
The R code is here on my RPubs page.
 | Nickolas on NYSE1996 Prices |
 | like this on NYSE1996 Prices |
 | image source on NYSE1996 Prices |
 | Over Here on NYSE1996 Prices |
 | stephenpeplow34 on Non-linear effects |
Spatial data and statistics 
