The other day I was discussing extreme global weather patterns with some of my colleagues, El Niño in particular. I started thinking that, with the help of a predictive model and our PMML Plug-in, we could use Datameer to predict the temperature of ocean currents, which is responsible for this type of climate phenomena.
To kick things off, let’s use a sample WeatherData set I have, which includes a variety of information like latitude, longitude, humidity components, etc. Initially, I go under Admin to the list of plug-ins where I will see a number of options that I can select from — we have quite a variety but you can always create a customized plug-in if you need something specific. Select the Universal PMML option (by Zementis) toward the bottom of the list, select configure and then choose your existing PMML file you want to apply to your data set (in this case it’s ElNino.pmml). Now you can go back to the Browser tab and add the WeatherData set to a new workbook.
Now I have a variety of columns that display my weather data in a structured, easy-to-read Excel-like format.
Once the data is in, I can apply my pre-existing PMML model. To do so, I start by selecting the next available column, which in this case is column G, and in the Formula Builder, select the PMML formula that I already have, and complete the arguments that I need (already available in my workbook under each column).
You will now see the results in column G. I can adjust the title to reflect what my predictive model is (StreamTemp) and round the decimals to a much more manageable number but in essence, I now know what the ocean current temperature would need to be when aligned with specific climate elements to create something as incredible as El Niño.
Check out the full tutorial below and be sure to download a trial version of Datameer to try this yourself: http://www.datameer.com/Datameer-trial.html?source=blog