A global humanitarian organization asked Elder Research to develop a customized training and collaborative modeling framework to empower their team to develop a baseline predictive model to optimize donor segmentation strategies.
in collaboration with the client's team, we initiated a process to establish a predictive modeling framework based on specific software tools such as RStudio, Packrat, and Git. A list of prerequisite reference material on using R for data preparation and modeling was provided to the client in advance on the on-site training. The one week onsite training and collaborative modeling engagement helped the marketing analytics team jump start their work in R by providing a blend of lecture and hands-on model building time using a data set developed on existing donors. Training topics included:
- Data Preparation Best Practices in R
- Source Version Control with Git
- Classification Techniques
- Model Selection
Under the guidance of Elder Research Data Scientists, the client's team developed a baseline multi-class classification model in R.
The framework streamlined the installation, setup, and management of necessary R package dependencies for the baseline model development and facilitated collaborative code development following open source best practices. The client was very satisfied with the quality of the training and insight provided.
"Elder Research provided our data science team with stellar customized R training and it was obvious they were well prepared with our data. The trainers were engaged, technically knowledgeable, and incredible communicators, which we thought would be a hard combination to find. We could sense their goal truly was to empower their clients and be unbiased data science partners, instead of trying to sell us a product."