About Algorithms I
In the previous article, we presented several different types of algorithms. In this one, we put some of them into use to address one of the climate change concerns.
So, what is the climate change problem we could, or would, address using data science? Towards that end, let us explore the opportunities of the use of data science into some of the different major climate change problems:
- Increased heat
- Drought
- Declined water supply
- Reduced agricultural yields
- Global temperature increases from human-made greenhouse-gases
In order for us to use data science to address climate change problems and concerns, we need to think in terms of what do we want to do, meaning what is the problem? What is the data science objective or the task to achieve? What data do we have access to? Which algorithm could help us? Towards that end, see the below table:
TABLE
| Problem/Concern | Objective/Task | Data | Algorithm |
|---|---|---|---|
| Increased heat | Predict heat | https://www.earthdata.nasa.gov/learn/pathfinders/disasters/extreme-heat-data-pathfinder | Regression analysis |
| Drought | Classify areas & Predict country-level drought | https://data.world/datasets/drought | Logistic Regression k-nearest neighbor, decision tress. Also ANN (deep learning for country prediction) |
| Decline in water supplies | Predict decline | https://www.eea.europa.eu/themes/water/dc | Regression analysis |
| Reduced agricultural yields | Segment geo | https://ourworldindata.org/crop-yields | Clustering |
| Human-made greenhouse gases | Predict & cluster | https://ourworldindata.org/co2-and-other-greenhouse-gas-emissions | Neural networks, support vector machines, clustering |
When you address a climate change problem or concern, it is important to think along those lines: the problem, the task, data, and algorithm. In the upcoming parts of the course, we will demonstrate experimental examples in order to show the “how” part of the above table could work, but you have to think of bringing your own problems and data to the above table, too.
Now, we run an algorithmic example using the dataset which is pre-installed into Orange. We encourage you to create an account to access the dataset. If you are unable to perform the exercise, we will describe the process below. Follow along, and find an example uploaded as an asset.
Steps to analyze the dataset
- Open Orange in your computer and go to File Menu
- Select New and add Add 5 widgets (File, Distances, Hierarchical Clustering, Data Table and Scatter Plot) to the workflow canvas > notice the connections in the figure
- Save as “Task3”
- Double-click on the File and Select the Iris dataset
- Double-click on the Distances and Select Euclidean
- Double-click on the Clustering widget to check the dendrogram
- Highlight the mixed-up data area
- Double-click both the clustering and scatter plot in order to check how the scatter plot changes with the area you highlight in the dendrogram. You should see the below screen:
Reach your personal and professional goals
Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.
Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.
