Repository for KUL course advanced analytics in Business
The input datasets can be found either datasets folder, or were to large to be included (assignment 2). A conda environment specifying all the packages and versions needed for the analyses per assignment can be found in the envs folder.
Note that assignment 2 does not have a conda environment, instead the notebook are meant to be ran in a docker environment that is based on a tensorflow-gpu-jupyter image.
- assignment1: Predictive modeling on tabular data
- assignment2: Image regression using CNNs
- assignment3: Streaming analytics on test data
- assignment4: graph analytics on social network data
Create a conda environment to house all the dependencies for assignment 1:
conda env create -f envs/assignment1.yaml -n advan_1
conda activate advan_1
jupyter lab
First create the docker container (ideally on a machine with a GPU available):
docker build -t tensorflow:1 envs/
docker run --gpus all -it -p 8888:8888 -v /path/to/repository:/tf --user 1000:1000 tensorflow:1
And run notebook assignment2 Note that the images should be available in the repository path under an 'images' directory.