PAL4C Labs | From Code to Clear Visuals: Data Science Fundamentals in Palantir Foundry

This PAL4C Labs walkthrough covers the data science fundamentals in Palantir Foundry. It shows how data scientists can clean data, analyze it with code, and turn the results into visuals anyone can understand. The project uses the provided patient datasets plus patient feedback stored as PDFs. 1. Building the pipeline The PDFs are processed and passed to an LLM, which returns each patient’s ID, a sentiment score and a summary of their feedback. The two patient datasets are joined, then combined with the LLM results. A final transform produces the clean dataset used for the rest of the analysis. 2. Analyzing the data with code The dataset is loaded into a Jupyter notebook and checked to confirm it imported correctly. Two types of charts are created using plotting tools and pandas. 3. Making it easy to read with Streamlit A Streamlit app turns the analysis into a clean, simple visual, so people don’t have to read code or scroll through raw data. At a glance, it shows how adverse side effects are spread across patient age groups. 4. Exploring the data lineage The lineage view shows how the source data flows through the pipeline into the final app. Built by PAL4C Labs, specialists in operational intelligence on Palantir Foundry and AIP.