Tuesday - Day 9
Accomplished:
Identified a bug in the present data compilation process and fixed it
Finished optimizing the data compilation code
Showed Quan how the new optimized code cleaner and faster
Paired with Quan to start building Logistic Regression model in MADlib
PDLtools is still a bottleneck, contingency plan is to just copy past code from pdltools github
Can bring up pdltools in Amy's meeting as its probably her team installing things on the CDE
Overview of the progress so far for Amy: First Week: - Investigated ways to integrate SAS models to score in GPDB, - Implemented scoring of the free-fall model while pairing with Victor - Walkthrough of various model workflows with Quan and Deepak
Second Week: - Presented the best practices & recommendations deck to Quan, Deepak and Victor - Deep dive into the frozen free fall model from Quan - Setting up logistics of network access, db access and required tools
Third Week: - Optimized the data compilation code - currently setting up code to run Logistic Regression in MADlib