Analysis with Andromeda
- An internet browser (this can be to access our running instance or to run your own: as explained here or using the Docker Container).
- A CSV with continuous, quantitative data, and at least one column with values unique across the dataset (this will be the label for the points).
- Note that non-numeric columns (other than label and URL values) will be dropped.
- [optional] URLs for all images identified in the CSV to display images next to the labels.
Sample CSVs are available under the datasets/ directory. Alternatively, a CSV can be retrieved from the "Fetch Data" Tab using an iNaturalist username with public observations. This also allows for direct integration of latloncover, which can be used to add landcover information to any CSV with columns indicating latitude and longitude of observations.
- Navigate to the Andromeda instance.
- Click “Choose File”, select your CSV, then click “Upload CSV”:
- Select the desired columns, then click “Visualize Data”.
- This will produce the following visualization using the sample dataset cartoonAnimals/Animal_Data_TaylorEtal.csv.
- Explore your data: regroup points or adjust the sliders to see how it impacts the projection. Be sure to select "Apply Moved Observations" or "Apply Slider Weights", respectively, to apply the adjustments to the entire dataset.
Moving observations (left), then applying those adjustments to the weights (right). Selecting "Apply Slider Weights" will re-arrange the points based on the displayed weights.