Note: You may use the data provided ONLY on your AWS virtual machine (VM). You are strictly prohibited from copying or sharing any of the datasets provided for this competition. Also note that your VM will be closed on 14 June 2020 at 23:59 GMT. All work on the VMs will be deleted at that time. It is critical that save a copy of your code (but not the data) elsewhere before that time.

The goal of this competition is to classify imagery at different locations, predicting whether a given location is within an informal settlement (1) or not (0). The imagery is 1.5m resolution SPOT imagery, covering the areas of interest. Both the imagery and the labels are from 2017.

Files available for download

• Train.csv - contains locations and the associated label (1 for informal settlements). The starter notebook shows an example of accessing imagery for a given location.
• Test.csv -is similar to train, but without the label column. For these locations, you must predict whether or not they fall within an informal settlement. Test data is sourced from a different province to the training data - make sure that your model is able to generalize well.
• Shapefiles - In addition to the Train.csv file, you are provided with a shapefile outlining informal settlements in the Gauteng province. You may use this data to generate additional training inputs for your model.
• SampleSubmission.csv - Shows the required submission format, with predicted labels for each location in the test set. The order of the rows does not matter, but the names of the IDs must be correct.