Empowering People
October 20, 2023
Empowering People
Score 9 out of 10
Vetted Review
Verified User
Overall Satisfaction with KNIME Analytics Platform
We use KNIME for three overlapping use cases. (1) With its drag and drop interface and visual management of software code it is a great tool for quick testing of concepts and building prototypes of data pipelines, machine learning solutions and data apps. With KNIME Analytics Platform, it is very fast to access and blend data from various sources including databases, APIs and flat files. KNIME's pre-built nodes cover a range of machine learning algorithms and associated procedures and where they fall short, its Python integration and shared components are likely to cover the gap. (2) As a free and accessible, but yet extremely powerful data tool KNIME Analytics Platform brings professional-level data processing and data science into the hands of anyone who wants to develop data skills beyond spreadsheets and BI systems. As the central analytics team, we can promote the tool to everyone whether they eventually became a user or not, without incremental cost. (3) The commercial product, KNIME Server/Business Hub enables turning the solutions developed in (1) and (2) into automated jobs and data apps accessible to anyone in the organisation.
- Easy access to powerful data wrangling capabilities to business users and citizen data scientists
- Simple management of complex analytical processes and user interfaces due to the visual workflow approach
- Straight forward integration with Python for additional capabilities
- Data Apps (KNIME Server/Business Hub) have the potential of moving self-service analytics and collaboration between business teams from creating and sharing BI dashboards into real applications with complex backends and rich user inputs
- The visualisation nodes that KNIME Analytics Platform offers out-of-the-box lack variety and configuration options to optimise their usability and looks for different use cases. However, the JavaScriptView and PythonView nodes together with the ability of using CSS styling should in principle provide boundless opportunities but are not necessarily accessible for those looking for a No Code/Low Code approach (also, the JavaScript nodes would benefit from similar package management approach to the Python integration). There are some user-driven developments and component nodes available on the KNIME Hub that improve the basic visualisation functionalities, but perhaps this is an area the KNIME team could also focus on with new nodes and components. One way of boosting development could be competitions for the user community focusing on visualisation approaches.
- Similarly, and related to the visualisation capabilities, the capabilities for creating Data Apps could be improved. More refined and intuitive user interaction within component views would require additional functionality. It would also be important to have more overall control of the app display and be able to create apps that do not follow the generic flow with standard [Next] and [Close] buttons, to disable the showing of the progress bar (which sometimes weirdly moves backwards rather than forwards) and to generate apps that can use the whole screen with fully customisable backgrounds. The objective should be to enable developing apps that the end-users will find intuitive and familiar based on their experience of mobile and other apps rather than expect users to adapt to certain idiosyncrasies of KNIME Apps.
- Intuitive data wrangling on KNIME Analytics Platform and deployment of data pipelines on the Server enabled us to insource previously outsourced BI development to a data science team residing within the business division, and subsequently exploring much more value adding solutions of ML/AI by combining our domain knowledge and technical skills.
- Alteryx and Dataiku DSS
As a commercial product Alteryx is more polished and can be even easier for a beginner, but KNIME beats Alteryx in functionality and performance. Dataiku takes the integration with Python and Git further than KNIME but isn't at the level of Alteryx and KNIME with its No Code/Low Code interface. In comparison to both Alteryx and Dataiku, KNIME is more versatile and significantly cheaper to deploy.
Do you think KNIME Analytics Platform delivers good value for the price?
Yes
Are you happy with KNIME Analytics Platform's feature set?
Yes
Did KNIME Analytics Platform live up to sales and marketing promises?
Yes
Did implementation of KNIME Analytics Platform go as expected?
Yes
Would you buy KNIME Analytics Platform again?
Yes
KNIME Analytics Platform Feature Ratings
Using KNIME Analytics Platform
10 - 3 (30%) of the users are in the data science/analytics team sitting in the business area. One person (10%) is the KNIME Server Admin at IT, with the rest of the users 6 (60%) being business users who use KNIME Desktop for personal automation and productivity (local workflows with lasting and widespread utility are deployed and managed on the KNIME Server by the central data science/analytics team). Additionally, training has been provided for all staff on moving from Excel to KNIME in the past. Around 40% of the current staff of the 40-strong business area have some experience of KNIME and more training on data wrangling is planned for the future.
4 - One person in the IT supports connectivity of KNIME to other in-house data infrastructure. A three person data science/analytics team with domain expertise and data and IT skills supports business users in utilising the tool in their day-to-day activities. Supporting of the business users requires understanding of their analytical objectives and processes, knowledge about the relevant data sources, the ability to convert spreadsheet-based processes into robust and streamlined pipelines and the skill to explain the conversion in layman terms.
- Personal productivity
- Orchestration of analytical procedures
- Productionising data science
- Exception management as an integral part of a data workflow
- The KNIME Business Hub may enable business users to share components and workflows
- With improvements to Data App functionalities, KNIME can have an important role in end-user facing interfaces
- Other user-facing apps than those running on KNIME can utilise the KNIME Server/Business Hub's API endpoint and use KNIME workflows as their backend
Evaluating KNIME Analytics Platform and Competitors
- Integration with Other Systems
- Ease of Use
KNIME enabled expanding self-service by business teams from consumption of dashboards and managing spreadsheets to repeatable data pipelines and machine learning applications.
I wouldn't change anything. There aren't other tools that would combine the ease of use, powerful functionality and cost-free nature of the KNIME Analytics Platform.
KNIME Analytics Platform Implementation
- Implemented in-house
Yes - We started by using individual installations of the open-source KNIME Desktop and after proving the value of the tool, the Server product was subscribed to.
Change management was a big part of the implementation and was well-handled - KNIME was the first tool in the organisation that would provide a database access to business teams and enable them to develop powerful data tools and apps independent of the IT department. This required us to partly redefine the roles and responsibilities of the business teams vis a vis the IT department, but this was all done in collaboration between the teams and solutions were found quickly.
- We encountered crashes of the KNIME Server when utilising certain memory-intensive Machine Learning nodes. It took longer than expected to fix the issue given the complexity of the Server product and limited Windows expertise at KNIME
KNIME Analytics Platform Support
Pros | Cons |
---|---|
Quick Resolution Good followup Problems get solved Kept well informed No escalation required Immediate help available Support understands my problem Support cares about my success Quick Initial Response | None |
We subscribe to KNIME Server which includes dedicated user support.
Using KNIME Analytics Platform
Pros | Cons |
---|---|
Like to use Relatively simple Easy to use Technical support not required Well integrated Convenient Feel confident using Familiar | None |
- Drag and drop data pipeline and machine learning development in general
- Reading data from Excel spreadsheets ignoring header rows or extra columns
- Integrated deployment of production solutions
- There are some legacy issues with nodes and packages developed at different times or by different teams that can have overlapping functionality and that may not work together even if they are seemingly functionally related
- The visualisation nodes are not always intuitive to use
Integrating KNIME Analytics Platform
- Azure SQL Server
- Anaconda Python
- Microsoft Excel
- File import/export
- API (e.g. SOAP or REST)
- Javascript widgets
- ETL tools