Introduction

Scenario: A regional distribution team is seeing rising late shipments and asks you, the new analyst, to find what supply-chain data is available to investigate.

Day 1 orients you inside Palantir Foundry and gets you fluent in finding supply-chain data before you analyze any of it. The regional distribution team is reporting rising late shipments, and your job as the new analyst is not yet to solve the problem but to map the terrain: what data exists, where it lives, how trustworthy it is, and who owns it. You will navigate the platform through the workspace sidebar and Quicksearch, locate the training Project that contains the relevant resources, and learn to read a dataset before you reason about it.

This matters because every later day of the disruption-response problem stands on the data you survey today. In Foundry, every dataset, report, and application is a 'resource', and resources live inside 'Projects' that form collaboration boundaries with Viewer, Editor, and Owner roles. Before trusting a number like 'late shipment rate', you must know the dataset's schema, when it last built, and which upstream sources feed it. Data Lineage and Dataset Preview are the two tools that turn an unfamiliar table into something you can defend.

By the end of the day you will produce a short 'data landscape brief' that the distribution team can act on: which datasets are in scope, what each one contains, how fresh and reliable it is, and who to contact when something looks wrong. This brief becomes the foundation for the investigation that runs through the rest of the curriculum.

Capability focus: Navigation, Projects/resources, Data Lineage, dataset preview · Artifact: Data landscape brief

Key concepts

Companion video

1. 1

   Find the training Project and core datasets

   In Foundry, all related supply-chain work is grouped into a 'Project', the container that holds resources and sets collaboration boundaries through Viewer, Editor, and Owner roles. Open the workspace navigation sidebar (Cmd/Ctrl+O) and use Quicksearch (Cmd/Ctrl+J) in full-results mode, or browse Files & Applications through Compass, to locate the training Project for the distribution team's data. Open it and note the folder structure and the core datasets inside, such as shipments, carriers, and inventory, since these are the resources your late-shipment investigation will draw on.

   not startedinstance check

   Run check

   Confirms all five supply-chain datasets resolve in the project.

2. 2

   Inspect the shipments dataset schema

   Dataset Preview lets you examine a dataset's contents without running any pipeline, showing the schema with each column and its data type, sample rows, and per-column statistics. Open the shipments dataset and use Preview to confirm the key columns you will need, such as shipment ID, promised delivery date, actual delivery date, carrier, and origin/destination, and verify their data types. Scan the column statistics for null counts, suspicious distributions, or outliers so you understand the quality of the late-shipment signal before you trust it.

   not startedinstance check

   Run check

   Confirms the shipments dataset exposes the columns you need for a late-shipment investigation.

3. 3

   Trace lineage, freshness, and ownership

   Data Lineage presents a graph of how datasets, transforms, and outputs connect, letting you explore upstream sources to learn where the shipments data originates and follow downstream consumers to see who depends on it. From the shipments dataset, open the lineage view, trace its upstream inputs, and use the build timeline to check when it last updated and whether it is flagged out-of-date, which tells you if rising late shipments reflect current reality or stale data. Check the resource's ownership and permissions so you know who to contact if the data looks wrong or you need additional access.

   not startedself-attested

   Mark complete

   Self-attested: you traced Data Lineage and noted freshness/ownership/quality risks.

4. 4

   Write the data landscape brief

   Synthesize what you found into a concise brief the distribution team can act on, drawing only on what Preview and Lineage actually showed rather than assumptions. For each in-scope dataset, record what it contains (key columns and types), its freshness (last build date and any out-of-date flags), its upstream sources, and its owner from the resource's access details. Reference each dataset by its name and RID so teammates can reliably locate the same resources, and flag any data-quality concerns that could distort the late-shipment analysis to come.

   not startedself-attested

   Mark complete

   Self-attested: your brief connects each dataset to an operational question.