Palantir Foundry Reviews

I am getting into the ontology space using Palantir Foundry. The primary use case is for developing a common business model that includes data, people, and processes, essentially describing how businesses operate. We are applying this model in the utilities sector.

Palantir Foundry offers the core capabilities of a digital twin enterprise. The digital twins concept allows bi-directional data and process integration between virtual and physical worlds. This integration enables what-if analysis, such as simulating tariff changes to plan responses. The low-code app development platform empowers any user to build applications using preferred visualization tools, while the AI platform allows for diverse model integration. Furthermore, Palantir Foundry is not limited to one LLM model, offering flexibility to build personalized models.

Our use cases are mostly related to data analytics. We are building some dashboards and ETL pipelines on the Palantir side. Palantir Foundry is a low-code/no-code platform with a user-friendly UI. It is better than Databricks, where you need to code. Palantir Foundry has better data lineage. However, Databricks also provides many features with Databricks Unity Catalog.

I like the data onboarding to Palantir Foundry and ETL creation.

I didn't use Foundry, but I went through some training, and my team became certified in it. When I left the company, there were probably 100 research projects that had been added to it. I did it project by project. Around 30 were completed. About 40 or 50 were in progress, while there were 20 more in the queue. You could reuse data and leverage data that had been imported. We imported lots of Epic data.

You needed permission to see the Epic data.  Someone with a research project approved by the institution could ask permission to join it with other data. In a relational world, you could say, "I'll give you database permissions, but I'll need to mask these columns that are based on those." It's similar to an SQL database.

People submitted their project requests to a project review committee. The capacity was limited because people needed to understand the platform, but I'm sure they have trained more people on it since then.

Foundry's data visualization is fantastic. If you had a solid data person sitting next to a researcher, you could do remarkable things with data visualization and transformation. 

The security is also excellent. It's highly granular, so the admins have a high degree of control, and there are many levels of security. That worked well. You won't have an EDC unless you put everything onto the platform because it is its own isolated thing.

We mainly used the solution to ingest our internal data. 

We have a client which uses it for insurance. The insurance company needed to calculate the premiums. In order to do that, we basically get our data from many regions, maybe New Zealand or Australia, Asia, and India. We focus on those regions. 

From those regions, get data in the form of Excel sheets or CSV files. We ingest those formats into the Palantir Foundry platform. And after ingesting, our data scientists upload the stream. Basically, they are trying to calculate the premiums on the basis of the data funneled over to them.

The best thing about the solution would be the data lineage. We use it to see the data sources in one place, including where the data is coming from and where it is going. The visibility is very easy. Right now, I'm working with another cloud platform and in that, and I don't see that option. I'm not able to easily know the sources. In Palantir, it was quite easy.

The solution is stable. 

It's scalable.

Palantir Foundry is being used for multiple hybrid cloud integrations in one of the services we provide for an existing US-based customer. It's all about getting together data from Azure and Amazon and then providing a hybrid platform through Palantir Foundry. We then provide the analytics or insights enablement for the customer.

Palantir Foundry is a robust platform that has really strong plugin connectors and provides features for real-time integration. The solution is more scalable and robust compared to other hyperscalers.

I use Palantir to build pipelines and ingest data from external sources like data lakes and bring that data from the data lake to the Palantir environment. Then, I use it to process data using some languages like Spark or SQL. The solution uses Spark for data processing. I can use Spark in the Palantir environment and create reports based on it.

The data lineage is great. It helps identify the flow of the data and the data pipeline, starting from the beginning to the end of the pipeline. 

I also found a contour helpful. It was also useful to do quick ad hoc data analysis and create reports.

Our primary use case is for data engineering and some data analysis, bringing in data from several sources and using data wrangling and data managing to support the reporting tools we have. We use the reporting apps for some of our basic reporting. We are customers of Palantir. 

The value of Palantir is that it provides everything in one tool. Previously, we'd need one tool for data cataloging, another as a data lake, and a third for data warehousing. Palantir has simplified that for us. The data analysis, data science and data engineering features are pretty easy to use.

Our company uses the solution to elaborate off PySpark. We run PySpark scripts for our products but use the solution to leverage the workflows that automatically run time bound and trigger bound. When a data trigger is there, the solution executes and the script starts running. 

Our scripts were long so they ran for two to three hours because they were generating some kind of commentary. The solution takes raw data as input, processes it, and converts it into a fine commentary for products.

More than 100 staff use the solution. 

The solution effectively processes all of our data within a single management platform. Previously, we had to use three different tools such as Hadoop to fetch data and leverage the use server. 

It is easy to map out a workflow and run trigger-based scripts without having to deploy to another server. This allows us to maintain an environment within the solution that goes directly to production.