Data Mining Curricular Unit of MDSAA at Nova IMS This is an ongoing process effort by:
Daniel Correia
Joana Rafael
Ricardo Santos
The goal is to perform donor segmentation of PVA dataset in the hope of developing targeted marketing strategies to incentivize lapsed donors to keep contributing to the non-profit association.
We tried 3 different clustering approaches and chose Self organizing maps as the best method for our approach. You will find the projects notebooks in the Projeto foldeer.
Each Notebook contains a portion of the project:
- DATA_CLEAN_EXPLOR
- KMEANS_HC
- GAUSSIAN_MIXTURE
- SOM_HC
- CLUSTER_CHARACTERIZATION
Our solution requires notebooks 1, 4 and 5. However. we keep notebooks 2 and 3 to showcase our tests.