State of DSML Platforms 2024

A week ago Gartner, the research and consulting company, released it's latest report on data science platforms. I'm tracking this research since 2020 and it's predecessor reports since early 2016. Before 2020, the Gartner Magic Quadrant for Data Science and Machine Learning Platforms was known as "Magic Quadrant for Advanced Analytics Platforms" and it's first release was 2014 February...

The Gartner Magic Quadrant is a research methodology and a very easy to understand graphical representation, a 2x2 matrix or a BCG matrix which was popularized by the Boston Consulting Group in the 1970s, used this time to evaluate and compare technology vendors based on two major aspects,

1. their ability to execute
2. their completeness of vision within the specific market.

When you segment platforms you get 4 regions;

• Leaders: well-positioned for future growth and direction
• Visionaries: possess innovative ideas but may not yet execute effectively.
• Niche Players: may lack broader innovation or outperforming capabilities.
• Challengers: may not demonstrate a clear understanding of future market.

From a short term benefit this this and similar research helps organizations make informed purchasing decisions about their technology investments.

But there is a longer term side as well, by looking and 2 or more consecutive years of reports you can understand how players in the market behave to new #innovations, how the adapt market changes and #finetune their #strategies to stay relevant in the game.

This is a better analysis if you are a vendor making a #longterm investment or if you are a strategy/innovation/product manager for one of the vendors planning your way towards leadership...

Very easy way to do this is to put two consecutive years reports on top of each other in transparent mode to see who "wins" (added into the quadrant or improves in execution and strategy towards leaders compared to last year), who "loses" (degrades in execution and/or strategy or removed completely out of the quadrant) and ties (confused so that some improvement in one dimension but degradation in another)

Here we go for this years interesting DSML report comparison;

Databricks is "the" clear winner form the perspective that they become "the best" in terms of ability to execute for the year 2024...

What's shocking is that Altair have sprinted from a very niche, "confused" "underperforming" player in 2021 to leaders section in 2024. This is due to two very smart investments;

• First is the acquisition of WPL, WORLD PROGRAMMING LIMITED which is the only platform who can run SAS code other than 赛仕软件 .
• Second is then a bold investment into Rapidminer platform, a well known #nocode analytics european player in DSML last year competing with Alteryx , KNIME and Dataiku . This literally catapulted Altair one of the longest one shot strides in DSML history...

Should note that cloud players Amazon Web Services (AWS) , Microsoft Azure & Google Cloud become the incumbent AI platforms for enterprises, not only with storage-> DB technologies, compute-> GPU's etc. but they started building very capable tools, for AutoML, forecasting, computer vision, squeezing the time to value for many use cases...

One notable mention is to Posit open-source data science platform, who is new to the quadrant.

There are also some players slipping due to very rapid innovation going on in the market...

Although they made many great additions to their IBM watsonx platform and their developments in #GenAI models they retreated to "challengers" from being a "leader" for many years... MathWorks in a worse situation dropping down to Niche players quadrant...

赛仕软件 and Alteryx are having a tough time due to their common strategy of

• thinking #GenAI #LLM's #VLM's are just a fad or "hype" and
• focusing on either legacy businesses like banking risk or insurance fraud in SAS's case and
• focusing on mostly structured #dataprep and #dataquality businesses like both in SAS and Alteryx.

However the money making machine in AI is gaining from #continuous labeling then learning and #seamless machine to human interactions with AI wether it's text based or voice based conversational interfaces or consuming realtime video streams. I predict companies who cannot build these capabilities immediately will struggle and become legacy pretty fast...

By the way we totally lost 三星SDS and Anaconda, Inc. is on the brink...

Please leave your opinions and predictions for next report!

All the best