Palantir Foundry Room for Improvement
MW
Mark Wozny
Manager, Data Governance at a healthcare company with 5,001-10,000 employees
The data lineage was challenging. It's hard to track data from the sources as it moves through stages. Informatica EDC can easily capture and report it because it talks to the metadata. This is generated across those various staging points.
It was hard to generalize that and put it into a catalog for people who you didn't know to reuse data. Maybe it's different now. That's the nice thing about Informatica: The catalog is reusable. Palantir is successful, and the institution loves them. I don't want to disparage it. I'm just speaking technically from an interoperability perspective.
View full review »Sometimes, when we used to ingest the data, it was not able to know the schema exactly. For example, if we had some semi-structured pile. If we uploaded that into Palantir, maybe sometimes it would not be able to find the correct schema which we were expecting at first.
The solution could use more online documentation for new users.
Technical support isn't very quick to help.
It's an expensive product.
View full review »KM
Keshav Mandal
Senior Analyst/ Customer Business and Insights Specialist at a tech services company with 501-1,000 employees
There is not a wide user base for the solution's online documentation so it is sometimes difficult to find answers. It is easy to find answers for code issues because Spark SQL and Python have wide user bases. There is a certain probability you won't find a solution-specific answer if you search for it. For example, there are certain errors that are specific to the solution. The more you use the solution, the more you understand it. The learning curve could be reduced with online documentation that includes the meaning of and troubleshooting for error codes.
Predefined code templates or informational prompts would help with writing syntaxes.
View full review »Buyer's Guide
Palantir Foundry
March 2024
Learn what your peers think about Palantir Foundry. Get advice and tips from experienced pros sharing their opinions. Updated: March 2024.
767,496 professionals have used our research since 2012.
The solution's pricing is high. Compared to other hyperscalers, Palantir Foundry is complex and not so user-intuitive. There could possibly be a little bit of overhead concerning the maintainability of the platform.
View full review »SB
Sai Arun Bandhakavi
Associate Vice President at a insurance company with 10,001+ employees
The application development aspect can be improved significantly and would make a difference. We use some third-party tools for reporting and it's a challenge for us to move data into the file system because Palantir is a closed environment and there are difficulties receiving data from external sources. There are some options in place for dealing with that but it's not sufficiently intuitive. I'd like the data exporting functionality to be as intuitive as the importing.
MC
Marco Cheung
Certified developer on Palantir Foundry
The frontend capabilities of Palantir Foundry could be further improved. You can build much faster than other products, but because of the speed, you lose certain capabilities.
Compared to Power BI, it has way more capabilities for editing, but it does not have fancy chart capabilities out-of-the-box as Power BI.
For future releases, I want the best of both worlds. On one hand, I would like more flexibility and connection to external libraries. On the other hand, I would like some of the wonderful out-of-the-box charting capabilities.
View full review »It doesn't have public documentation like AWS or Azure.
For data modeling, there were tools like Object Explorer and front-end applications. I found it difficult. Even though I don't use it extensively for my project, it was difficult to understand the functionality of those apps. The ones that I used were normal data lineage code tools and had no data preparation. That said, for others, there was a slight application for a front end. It was a bit difficult. If Palantir can add more documentation to this part, including how to use it and how to create a simple application, it would be helpful.
It would be helpful to build applications based on Azure functions or web apps in Palantir Foundry. If that kind of capability was available there, it could be nice.
View full review »RD
reviewer2031192
Data Scientist at a financial services firm with 10,001+ employees
Some error messages can be very cryptic when things don't run according to plan or ontology. As a new user, this creates a bit of a black box because you don't know what issue causes a particular type of error. It takes a lot of time or trial and error to understand the software.
The stats feature that details occupied memory should be available to the developer role. As developers, we cannot see in real time how much memory is being used or which flow is being executed.
View full review »The workflow could be improved. Although it works rather seamlessly, the workflow too complicated sometimes. Maybe they can reduce the complexity of the workflow. It could be more modularized in the future.
The performance of the engine could be better.
View full review »AD
reviewer2038020
Data Engineer at a manufacturing company with 10,001+ employees
Computing is very expensive. If you want to create new models on specific data sets, computing that is quite costly.
Python's current setup within Palantir is very limiting. I would like to have more freedom to use Python without limitations.
CD
reviewer2022087
Sr Consultant at a consultancy with 501-1,000 employees
Palantir Foundry is marketed as an auto-magical tool that can take over your existing system via artificial intelligence and build the building blocks for it. But it requires a lot of manual work and is very time-consuming to get to a functional point. Therefore, the setup of the system needs to be improved.
View full review »PZ
reviewer1379157
Associate - Inhouse Consulting at a pharma/biotech company with 10,001+ employees
They do not have a data center in Europe, and we have lots of personally identifiable information in our dataset that needs to be hosted by a third-party data center like Amazon or Microsoft Azure.
There are some issues with scalability because when we are using a really large dataset, the system is rather slow.
The performance can be improved. It would make our life a lot easier if it were as fast as Google Cloud. The GCP is unmatchable in terms of the speed at the moment.
From a user perspective, it would be nice to have a preview of what the data is looking like. As it is now, you can see the schema but not the actual data. For example, they can see the different columns but they don't know what's there. If they could inspect the first few hundred columns of data then they would have an idea of what they are dealing with.
View full review »SB
Shikha Bansal
Senior Manager at freelancer
The one area where improvement could be made is the cost of the solution which is quite expensive.
View full review »
Buyer's Guide
Palantir Foundry
March 2024
Learn what your peers think about Palantir Foundry. Get advice and tips from experienced pros sharing their opinions. Updated: March 2024.
767,496 professionals have used our research since 2012.